How collaboration arrises and why it fails

What if the visual brain does not process information through a single hierarchy but runs multiple parallel systems that complete their tasks at different times? Neuroscientist Semir Zeki challenges the textbook model of visual processing, arguing that asynchronous operations across parallel pathways, not sequential stages through V1, are the fundamental organizing principle of visual perception. Subscribe for more from the Convergent Science Network podcast series. Semir Zeki joins Paul Verschure and Tony Prescott at the BCBT summer school to present an alternative architecture for the visual brain built on four key findings that the standard model fails to accommodate. First, V1 is not the sole gateway to visual cortex , direct projections from the LGN and pulvinar reach specialized visual areas independently. Second, conscious visual experience can occur without V1, either as preprocessor or postprocessor. Third, different visual attributes are perceived at different times: color is seen before form, and form before motion, with gaps of up to 80 milliseconds. Fourth, the brain operates asynchronously, and no current computational theory adequately accounts for this. The discussion examines what determines which of the multiple anatomical hierarchies takes precedence at any given moment. Zeki proposes that the answer is task- and stimulus-dependent: the same physical substrate supports many possible functional hierarchies, dynamically configured according to what the brain needs to process. He presents evidence that signals from simple geometric elements reach both V1 and high-level face and house areas at identical latencies, challenging the assumption that complex object recognition is built exclusively from oriented line detectors in V1. The conversation also addresses the Gestalt principle , that a face may be recognized as a whole before its component features are analyzed , and why this demands rethinking the building-block model of visual processing. Key topics include the role of the pulvinar in attentional modulation, why perceptual latency hierarchies do not match physiological latency hierarchies, the relationship between fiber diameter and processing speed, Zeki's advocacy for genuine interdisciplinarity between neuroscience and philosophy, and what neuroaesthetics reveals about the brain's knowledge-acquisition function. Part of the Convergent Science Network podcast series from the BCBT Summer School.

Can a rat perform the same perceptual decision-making tasks that were once thought to require a primate brain? Neuroscientist Matthew Diamond explains how rats trained on complex vibrotactile comparisons reveal fundamental principles of evidence accumulation, working memory, and sensory coding , and why individual differences between rats rival those between humans. Subscribe for more from the Convergent Science Network podcast series. Matthew Diamond joins Paul Verschure and Tony Prescott at the BCBT summer school to present his laboratory's work on whisker-mediated decision-making in rats. Using a paradigm in which rats compare two vibrotactile stimuli separated by a delay, Diamond's team has shown that rats can perform parametric comparisons of stimulus intensity and duration , tasks previously considered beyond rodent capability. The results demonstrate that rats accumulate evidence over time from stochastic stimuli, improving performance with longer stimulus durations, consistent with optimal evidence integration. The discussion distinguishes between two modes of whisker sensing: receptive sensing, where the animal holds its whiskers still to collect an externally delivered vibration, and generative sensing, where the animal actively creates stimulation through its own whisking movements. Diamond argues both are forms of active sensing, since even in the receptive case the animal actively controls whisker state to optimize signal collection. The conversation explores how rats and humans compare on psychometric performance , on average humans outperform rats, but the distributions overlap substantially, with the best rats exceeding the worst human subjects. A key finding is that stimulus intensity and duration combine through summation rather than multiplication, suggesting the brain adds rather than multiplies evidence from these two dimensions. The discussion also addresses why rats show higher lapse rates than humans , possibly reflecting an evolved strategy of continuously exploring whether task contingencies have changed, rather than exploiting a known rule. Diamond explains how these rodent studies complement primate research by revealing how a simpler brain with fewer cortical modules can accomplish similar computations through different circuit architectures. Key topics include parametric versus categorical decision-making, evidence accumulation in stochastic environments, cross-modal comparison between auditory and tactile stimuli, individual variability in rat cognition, and what working memory in rats reveals about prefrontal cortex homology. Part of the Convergent Science Network podcast series from the BCBT Summer School.

How does the brain build itself from a handful of precursor cells into billions of neurons organized in precise layers? Developmental neurobiologist Stephen Noctor explains the remarkable choreography of cell division, migration, and differentiation in the ventricular zone , where future neurons bounce up and down like yo-yos before embarking on journeys equivalent to climbing four Empire State Buildings. Subscribe for more from the Convergent Science Network podcast series. Stephen Noctor joins Paul Verschure and Tony Prescott at the BCBT summer school to describe his research on the precursor cells that generate the cerebral cortex. Using fluorescent labeling and time-lapse imaging in rat brain slices, Noctor has captured the movements of individual precursor cells as they undergo interkinetic nuclear migration , rapidly descending to the ventricular surface to divide, then slowly rising back through the ventricular zone in a process that may be largely passive. His movies reveal surprising behaviors: cells that pause, reverse direction, and emit transient processes that may serve as feedback conduits during migration. The discussion traces cortical construction from its earliest stages. Excitatory neurons are generated in the ventricular and subventricular zones and migrate radially outward, while inhibitory interneurons originate in the ganglionic eminences and travel tangentially. Noctor estimates fewer than ten distinct precursor cell types, which become progressively restricted in their output as development proceeds. He describes the inside-out lamination of the cortex, where later-born neurons migrate past earlier-born ones to settle just beneath the marginal zone , a process dependent on the signaling molecule reelin, without which the cortex inverts. Key topics include the orientation of cell division planes and what they reveal about fate determination, the role of radial glial fibers as scaffolds for migration, why the human brain generates roughly five billion cortical neurons over the course of pregnancy, the forgotten discoveries of Frederick Sauer from 1935, and how understanding normal development establishes a foundation for investigating neurodevelopmental disorders. Part of the Convergent Science Network podcast series from the BCBT Summer School.

How does the brain transform a fleeting touch on the fingertip into a deliberate decision seconds later? Neurophysiologist Ranulfo Romo explains how sensory representations are maintained, transformed, and compared across cortical areas , revealing the slow, parametric neural code that bridges perception and decision-making. Subscribe for more from the Convergent Science Network podcast series. Ranulfo Romo joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss decades of work tracing how somatosensory signals travel from primary cortex to frontal decision-making areas in the primate brain. Using a vibrotactile discrimination task in which monkeys compare two temporally separated stimuli, Romo has mapped the transformation of sensory information at each stage , from faithful isometric representations in S1 arriving within 25 milliseconds, to slowly ramping parametric codes in prefrontal and premotor areas emerging around 180 milliseconds. The discussion addresses why Romo insists on unimodal processing in primary sensory cortex, a position he has tested for over a decade against the competing multimodal hypothesis. He argues that the neural doctrine , the idea that cortical territories are defined by their thalamic inputs , still holds, and that the key scientific question is how a sensory representation is progressively transformed as it passes through successive cortical areas, each treating the signal differently before passing it on. During the delay period between stimuli, frontal neurons maintain a ramping activity that preserves the stimulus parameter while discarding irrelevant features , a process Romo links to the ancient philosophical tradition from Democritus and Epicurus about how the brain generates internal representations of the external world. Key topics include the distinction between fast sensory and slow cognitive processing systems, the role of neuromodulators in bridging these timescales, why decision-making is context-dependent and sometimes unconscious, the relationship between Libet's conscious awareness timing and primate neurophysiology, and the challenge of procrastination as a decision-making phenomenon. Part of the Convergent Science Network podcast series from the BCBT Summer School.

What if the cerebellum is not just a motor structure but a universal learning machine wired to the entire frontal lobe? Neuroscientist Narender Ramnani explains how the anatomy of cortico-cerebellar loops forces us to rethink the cerebellum's role , from fine-tuning movements to supporting rule learning, cognitive error processing, and the transition from deliberate to habitual behavior. Subscribe for more from the Convergent Science Network podcast series. Narender Ramnani joins Paul Verschure and Tony Prescott at the BCBT summer school to present evidence that the cerebellum communicates not only with motor cortex but with diverse regions of the prefrontal and parietal cortex through closed anatomical loops. He traces this insight to work by Peter Strick and by Schmahmann and Pandya, which revealed that the cerebellum's connectivity is far broader than the classical motor view suggests. If the cerebellar microcircuit is computationally uniform , the same Marr-Albus learning architecture repeated across the structure , then the same transform applied to motor inputs should also apply to cognitive inputs arriving from prefrontal cortex. The discussion digs into the critical question of error signals. In classical conditioning, the inferior olive delivers a clear teaching signal. But what serves as the error signal for prefrontal-cerebellar loops? Ramnani presents anatomical evidence for at least two routes: dopaminergic projections from the VTA that send collaterals directly to cerebellar Purkinje cells, and prefrontal projections from the anterior cingulate cortex that reach the inferior olive. He also describes fMRI evidence showing that cerebellar activity during instrumental rule learning mirrors the Purkinje cell pause seen in classical conditioning , a decrease in BOLD signal consistent with reduced Purkinje cell firing during learning. Key topics include the modular independence of cerebellar loops, why cerebellar modules must communicate through neocortex rather than internally, the system-one versus system-two distinction in habit formation, how to interpret fMRI signals in the cerebellum, and the challenge of building computational models that capture these cortico-cerebellar interactions. Part of the Convergent Science Network podcast series from the BCBT Summer School.

What if the twitches you see in a sleeping infant are not remnants of dreams but a systematic self-calibration process for the developing motor system? Developmental neuroscientist Mark Blumberg explains how REM sleep twitching may serve as the brain's sonar , pinging muscles one at a time and listening for the sensory feedback that bootstraps the body map. Subscribe for more from the Convergent Science Network podcast series. Mark Blumberg joins Paul Verschure and Tony Prescott at the BCBT summer school to present his research on the relationship between sleep, twitching, and sensorimotor development in infant rats. During active (REM) sleep, neonatal rats produce highly discrete myoclonic twitches , brief activations of individual joints occurring against a background of low muscle tone. Blumberg argues these are not random byproducts but structured motor events with a high signal-to-noise ratio, ideally suited for the nervous system to map its own body. The sensory consequences of each twitch cascade through the brain in ways that wake movements do not, because during wakefulness a corollary discharge mechanism gates reafferent signals. The discussion traces the developmental trajectory of twitching from spinal-cord-driven activity in the fetus through brainstem contributions in the neonate, showing how the system becomes progressively more complex. Blumberg presents evidence that multi-joint twitches develop spatio-temporal structure over the first postnatal week, with certain movement combinations becoming more frequent and more organized , suggesting a selectionist process that shapes which movement patterns persist. He challenges the concept of motor primitives from a developmental systems perspective, arguing that no aspect of the motor system comes for free and that what appears primitive at one level is always a developmental product at another. Key topics include the distinction between REM sleep twitching and wake movement, how corollary discharge develops in early life, why the cortex appears uninvolved in producing twitches at early ages, the relationship between twitching and joint maintenance, and what the developmental perspective offers to robotics and motor control theory. Part of the Convergent Science Network podcast series from the BCBT Summer School.

How does the brain build a map of three-dimensional space when a full volumetric representation would be prohibitively expensive? Neuroscientist Kate Jeffery explains why the rat navigation system appears to favor flat maps stitched together into a mosaic , and what this reveals about the evolutionary trade-offs shaping spatial cognition. Subscribe for more from the Convergent Science Network podcast series. Kate Jeffery joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss her research on how place cells, grid cells, and head direction cells handle the vertical dimension. Her laboratory has found that grid cells, which fire in periodic hexagonal patterns on flat surfaces, do not produce the same metric structure in the vertical plane. On a pegboard where rats move horizontally at different heights, grid fields extend into strips rather than grids. On a climbing wall where the body is parallel to the surface, something more grid-like appears. The implication is that the system maps space relative to the plane of the animal's body rather than constructing a universal three-dimensional coordinate frame. The discussion addresses what this means for models of spatial cognition. Jeffery proposes a multi-planar model in which the brain tiles complex three-dimensional environments with locally two-dimensional map fragments, linked by some coarser three-dimensional information. She explains why this is an efficient evolutionary solution: a full 3D map would require vastly more neural resources, while a patchwork of flat maps supplemented with elevation cues handles most real-world navigation demands. The conversation also explores how the head direction system might cope with three dimensions , whether through a spherical attractor, three orthogonal ring attractors, or a simpler scheme that just tracks yaw on whatever surface the animal occupies. Key topics include the relationship between grid cells and contextual cues, the developmental sequence of spatial cell types, the influence of deep learning on thinking about modularity in the brain, and the practical constraints that ecology imposes on neural representations of space. Part of the Convergent Science Network podcast series from the BCBT Summer School.

What if the brain organizes thought not as a continuous stream but as a series of discrete packets, timed by nested brain oscillations? Neuroscientist John Lisman explains how theta and gamma rhythms work together to chunk information into ordered sequences , a coding scheme he proposed 20 years ago that recent experimental breakthroughs have finally confirmed. Subscribe for more from the Convergent Science Network podcast series. John Lisman joins Paul Verschure and Tony Prescott at the BCBT summer school to revisit his influential theta-gamma coding hypothesis, first published with Idiart two decades earlier. The core idea is that within each cycle of the slower theta oscillation (roughly 5–15 Hz), the brain fits approximately six or seven discrete gamma cycles (30–100 Hz), and each gamma cycle carries a distinct piece of information. In the hippocampus, this means different spatial locations are represented at different gamma phases within a single theta cycle , not as a continuous signal, but as an ordered, discretized sequence. The discussion explores what recent data from Foster and colleagues has added to this picture: direct evidence that hippocampal representations jump between discrete positions in space, locked to successive gamma cycles, confirming that the phase code is genuinely discrete rather than continuous. Lisman argues this amounts to a multi-part message delivered in under 100 milliseconds , a compressed movie of a navigational path that downstream structures like the basal ganglia could evaluate for costs and benefits during decision-making. The conversation also tackles deeper questions about whether the brain operates with anything resembling a clock cycle, how pattern completion can occur within a single gamma window, and why the irregularity of gamma timing does not undermine the coding scheme. Lisman, Verschure, and Prescott debate the relationship between episodic and statistical memory, the computational parallels to digital processing, and whether oscillatory codes represent a fundamental organizational principle or just one of many strategies the brain employs. Key topics include the theta-gamma nesting hypothesis, discrete phase coding in hippocampus, working memory capacity, attractor dynamics within gamma cycles, decision-making via sequential replay, and the role of brain oscillations in structuring cognition. Part of the Convergent Science Network podcast series from the BCBT Summer School.

If the adult brain cannot change, how did you learn anything after childhood? Neuroscientist Greg Recanzone revisits the revolution in adult cortical plasticity , from the landmark digit amputation experiments to his own work showing that perceptual training reshapes somatosensory maps through mechanisms fundamentally different from developmental critical periods. Subscribe for more from the Convergent Science Network podcast series. Greg Recanzone joins Paul Verschure and Tony Prescott at the BCBT summer school to tell the story of how adult cortical plasticity went from heresy to established fact. Beginning with Mike Merzenich and Jon Kaas's digit amputation studies in monkeys, Recanzone describes how the somatosensory map in area 3b completely reorganized to look like a normal four-fingered monkey , not just filling in a gap, but rebuilding topographic order. This was the key insight: receptive fields are dynamic, continuously adjusting synaptic weights relative to neighboring neurons. The consensus that emerged distinguishes developmental plasticity, which involves anatomical rewiring, from adult plasticity, which operates through synaptic weight changes and modulation of inhibition. The discussion then turns to Recanzone's own experiments training monkeys on a vibrotactile frequency discrimination task. The trained skin showed expanded cortical representation, enlarged receptive fields, and, most importantly, dramatically tighter temporal fidelity across the neuronal population. Individual neurons responded no better than untrained controls, but the trained population locked their responses to each stimulus cycle with far less variability, producing a louder and cleaner signal. This enhancement depended critically on task engagement and reward: passive stimulation with identical physical input produced no comparable changes, confirming that neuromodulatory signals gated by attention and reinforcement are essential for adult plasticity. Key topics include why Merzenich and Kaas faced years of resistance to their plasticity findings, how the reorganization following digit amputation differs from visual and auditory cortex lesion effects, why receptive field enlargement during training reflects Hebbian co-activation rather than task demands, what the role of neuromodulators like acetylcholine and dopamine is in gating cortical map changes, how Mike Kilgard's basal forebrain stimulation experiments confirmed that neuromodulation alone can drive map reorganization, and what the practical limits of adult cortical plasticity are for rehabilitation and skill learning. Part of the Convergent Science Network podcast series from the BCBT Summer School.

How does the brain build an internal map of space , and what happens when that map is slightly wrong? Nobel laureate Edvard Moser describes the discovery of grid cells, their modular organization, and the surprising geometric distortions that reveal how the brain calibrates its spatial metric against the physical world. Subscribe for more from the Convergent Science Network podcast series. Edvard Moser joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss his research on the neural basis of spatial navigation. The conversation traces the path from hippocampal place cells to the discovery of grid cells in the medial entorhinal cortex , neurons that fire in strikingly regular hexagonal patterns as an animal moves through space. Moser explains how targeting electrodes to a more dorsal region of entorhinal cortex, guided by neuroanatomist Menno Witter's expertise on hippocampal connectivity, revealed spatial signals that previous studies had missed simply because they recorded in regions where grid spacing was too large for standard-sized environments. The discussion explores the key properties of grid cells and their organization into discrete modules , clusters of cells with rigidly preserved firing relationships across different environments. Within each module, cells maintain consistent phase offsets, orientations, and spatial scales, providing a reusable metric framework that does not need to be rebuilt for every new environment. Moser describes how grid cells depend on speed and direction inputs for path integration but require continuous calibration against external sensory cues, particularly visual landmarks, to prevent cumulative drift errors. Border cells, head direction cells, and speed cells form a local circuit ecosystem that supports and anchors the grid representation. Key topics include how grid cells were discovered and why earlier studies missed them, what modular organization means for generating unique position codes from redundant grid patterns, how border cells anchor and distort grid patterns near environmental boundaries, why grid axes are offset by 7.5 degrees from wall orientations due to a shearing process, how hippocampal place cells create distinct orthogonal maps for different environments from the rigid grid cell input, and what the lateral entorhinal cortex contributes beyond spatial information to hippocampal representations. Part of the Convergent Science Network podcast series from the BCBT Summer School.

Can stress experienced by a mother rat change the brains of her great-grandchildren , and what does that tell us about how early experience shapes human development? Neuroscientist Brian Kolb presents evidence that epigenetic effects of stress, tactile stimulation, and drugs of abuse persist across at least four generations, with profound implications for understanding literacy, cognitive development, and public health. Subscribe for more from the Convergent Science Network podcast series. Brian Kolb joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss how experience interacts with gene expression to reshape brain circuits and behavior across generations. His research methodology follows a systematic pipeline: first identify behavioral changes, then locate synaptic reorganization in the brain using Golgi staining, then drill down to gene expression changes using methylation analysis and gene chip arrays. Using this approach, Kolb demonstrates that prenatal stress in rats produces increased anxiety, impaired motor and cognitive skills, and measurable changes in prefrontal cortex synaptic organization , effects that persist through at least four generations and can even be transmitted indirectly through a stressed animal's communication with its unstressed mate. The discussion bridges animal research and human development through a compelling analysis of vocabulary acquisition. Children in higher socioeconomic status families are exposed to roughly one million more words by age three, largely through serve-and-return social interaction, setting them on a trajectory that eight years of schooling fails to reverse. Kolb presents evidence from Cuba, South Carolina, and Sweden showing that early intervention programs that pour resources into the first three years of life produce dramatic improvements in literacy and cognitive skills, regardless of the population's baseline. He connects this to his animal work through the mechanism of tactile stimulation, which releases FGF2 and produces widespread synaptic changes and enhanced cognitive abilities in offspring. Key topics include how stress, drugs, and tactile stimulation each leave distinct epigenetic footprints in the brain, why bystander stress transmitted through ultrasonic vocalizations affects offspring development, how early stress may inoculate against later stressors at the cost of reduced cognitive capacity, what the Barker hypothesis predicts about adaptive responses to dangerous environments, and why the first three years of life represent a critical window that determines lifelong cognitive trajectories. Part of the Convergent Science Network podcast series from the BCBT Summer School.

How does an animal with no skeleton, no somatotopic brain map, and eight arms containing more neurons than its central brain manage to produce precise, goal-directed movements? Neuroscientist Benny Hochner reveals how the octopus solves the seemingly impossible problem of controlling a soft body with infinite degrees of freedom. Subscribe for more from the Convergent Science Network podcast series. Benny Hochner joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss his research on motor control and learning in the octopus , an animal he describes as the most intelligent invertebrate and a remarkable case study in convergent and divergent evolution. With half a billion neurons, most distributed across its eight arms rather than centralized in the brain, the octopus has evolved a radically different solution to motor control than vertebrates. For reaching movements, it reduces its theoretically infinite degrees of freedom to just three by propagating a wave of muscle stiffening along the arm, creating a simple but effective motor program that can be generated even in a completely severed arm. The discussion explores the hierarchical organization of the octopus nervous system, from autonomous arm reflexes to coordinated whole-body behavior. A severed arm can still grasp food and pass it along its suckers toward where the mouth would be. The central brain appears to encode motor programs rather than body maps , no somatotopic organization has been found for either motor commands or sensory processing. Remarkably, tactile discrimination learned with one arm generalizes to all others, confirming central involvement in learning but not in arm-specific representation. Hochner also describes convergent findings in learning and memory: the octopus vertical lobe resembles the mammalian hippocampus in structure and exhibits robust activity-dependent long-term potentiation, though mediated by molecular mechanisms modified from simpler molluscan ancestors. Key topics include why the octopus is scientifically important as an independently evolved intelligent invertebrate, how muscular hydrostats solve the degrees-of-freedom problem through embedded motor primitives, why no body map exists in the octopus central brain, how the fetching movement creates a temporary articulated structure from a boneless arm, what the vertical lobe reveals about convergent evolution of learning mechanisms, and how the octopus challenges conventional assumptions about the necessity of body representation for coordinated action. Part of the Convergent Science Network podcast series from the BCBT Summer School.

Why has the basic architecture of the vertebrate brain remained essentially unchanged for 450 million years , and is that a constraint or an optimal design? Evolutionary neuroscientist Barbara Finlay presents evidence that mammalian brain development follows a remarkably conserved nonlinear timetable, transformable across species with 99 percent accuracy by turning a single dial. Subscribe for more from the Convergent Science Network podcast series. Barbara Finlay joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss the principles underlying brain evolution, drawing on her translatingtime.net database spanning 18 mammalian species. Her central finding is striking: the developmental schedule of the brain, from the birth of the first neurons to the onset of behavior, can be transformed from mouse to cat to monkey to human by a single nonlinear function with extraordinary precision. This conservation extends to remarkably specific events, including when Purkinje cells are born and when layer four cortical neurons are generated. The discussion explores whether this invariance represents a developmental constraint or an actively defended optimal design. Finlay argues for the latter, noting that 450 million years of evolution have preserved this architecture across radical changes in niche , from water to land to air and back. She identifies four core learning engines present in the earliest vertebrates , cortical association, hippocampal memory, basal ganglia reinforcement learning, and cerebellar optimization , and proposes that this combination may explain the explosive success of vertebrates. The conversation also examines how the relative sizes of brain structures trade off, particularly the inverse relationship between isocortex and olfactory bulb, which appears to be mediated by timing shifts in neurogenesis. Key topics include how a nonlinear developmental timetable predicts brain structure timing across all mammals, why the lateral edges of the embryonic brain produce the most variable and plastic structures, what Paul Katz's catalog of swimming marine mollusks reveals about the limits of evolvability, how critical periods may be self-initiated by appropriate input rather than fixed in time, and why the allocation of neural resources between sensory modalities follows predictable patterns shaped by both development and ecological niche. Part of the Convergent Science Network podcast series from the BCBT Summer School.

Is the brain a dynamical system, an information processor, or a prediction machine , and does it even matter which label we choose? Computational scientist Randall Beer argues that these are not competing theories but complementary mathematical lenses, and that real progress requires building theory around carefully analyzed toy models rather than debating metaphors. Subscribe for more from the Convergent Science Network podcast series. Randall Beer joins Paul Verschure and Tony Prescott at the BCBT summer school to present his approach to understanding brain, body, and environment as coupled dynamical systems. Beer makes a sharp epistemological argument: statements like "the brain is a dynamical system" or "the brain is an information processor" are not testable theories but pre-theoretical intuitions, each backed by a body of mathematics that serves as a lens for examining neural systems. No experiment could definitively prove or disprove any of them. What matters is the utility of each lens for generating insight, and Beer advocates maintaining a toolkit of multiple mathematical languages rather than committing to any single framework. The discussion centers on Beer's detailed analysis of a minimal agent performing relational categorization , distinguishing the relative size of two falling objects. Using both dynamical systems theory and information theory applied to the same evolved neural controller, Beer demonstrates that each lens reveals complementary features invisible to the other. Dynamical analysis highlights bifurcations, transient manifolds, and the role of sensor discontinuities, while information-theoretic analysis reveals which combinations of system elements carry the most relevant information at each moment. The invariant pattern across many evolved solutions is a transient manifold that gets spread into a sheet and then sliced by a bifurcation into a decision. Key topics include why brain-body-environment should be the unit of analysis rather than the brain alone, how toy models in the tradition of Galileo's frictionless planes can build fundamental theory, what the difference is between ontological and epistemological claims about neural computation, why dynamical systems theory and information theory are complementary rather than competing, and how Beer plans to extend these analytical tools to the biological nervous system of C. elegans. Part of the Convergent Science Network podcast series from the BCBT Summer School.

Did birds and mammals independently evolve the same brain circuit , and what does that mean for how we define homology? Developmental neurobiologist Zoltan Molnar presents evidence that avian and mammalian forebrains share strikingly similar gene expression patterns and functional properties despite arising from different parts of the embryonic brain. Subscribe for more from the Convergent Science Network podcast series. Zoltan Molnar joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss the evolution and development of the neocortex, tracing the question back to Thomas Willis's 1664 observation that the cerebral cortex is disproportionately enlarged in humans. Molnar argues that while the cortex is clearly central to higher cognitive function, understanding its evolution requires confronting one of the thorniest problems in evolutionary biology: the relationship between the mammalian neocortex and the avian dorsal ventricular ridge, structures that show convergent gene expression, similar electrophysiological properties, and comparable circuit organization, yet develop from different regions of the embryonic telencephalon. The discussion produces a spirited debate about the meaning of homology. Molnar insists on a developmental definition , structures are homologous only if they derive from the same part of the neuroepithelium , and presents lineage-tracing evidence that mammalian layer 4 neurons and avian nidopallium neurons originate from distinct progenitor populations. Verschure and Prescott push back, arguing that convergent gene expression and functional equivalence in the adult brain may warrant a broader evolutionary definition. The conversation also covers how highly conserved homeobox genes mark early brain segments identically across vertebrates, how thalamic inputs may drive convergent differentiation in recipient cells regardless of their developmental origin, and what Harvey Karten's equivalent circuit hypothesis means in light of modern transcriptomic data. Key topics include why neocortex is the key structure for understanding brain evolution, how conserved developmental programs constrain but do not fully determine adult brain organization, what the reeler mutant reveals about the robustness of cortical self-organization, how sauropsids and mammals enlarged different parts of the forebrain, and why the debate over homology versus convergence remains unresolved despite decades of comparative data. Part of the Convergent Science Network podcast series from the BCBT Summer School.

Can you grow an organ inside a computer , and would it teach you something biology alone cannot? Computational biologist Yaki Setty describes how agent-based models of stem cell development can reconstruct the pancreas, the C. elegans gonad, and even parts of the brain from first principles, revealing emergent properties that no single experiment could predict. Subscribe for more from the Convergent Science Network podcast series. Yaki Setty joins Paul Verschure and Tony Prescott at the BCBT summer school to present his approach to synthetic organ development using autonomous agent-based modeling. Each cell in the simulation is defined by biologically justified state diagrams , covering differentiation, proliferation, movement, and environmental sensing , with every parameter traceable to published experimental data. The environment is modeled as a three-dimensional grid of voxels containing chemical gradients governed by differential equations, and cells interact with this environment and with each other to produce emergent organ structures. The discussion walks through three applications of increasing complexity. The pancreas model, with over 150 cell states, reproduces the characteristic cauliflower-like morphology of pancreatic tissue and demonstrates how blood vessel scaffolding guides cell aggregation. The C. elegans gonad model achieves quantitative predictions about cell numbers, zone lengths, and cell cycle ratios with far fewer states, validated against experimental measurements within weeks rather than years. The conversation also touches on extending these methods to neural development, where the same platform and principles apply but the complexity of cell types and connectivity presents new challenges. Key topics include how autonomous agent models differ from conventional computational approaches, why all available biological data should be incorporated rather than held back for testing, how mutations serve as the primary validation strategy for these models, what the relationship is between stem cell stemness and differentiation potential, why morphological benchmarks like cauliflower structure are difficult to quantify rigorously, and how these simulations could eventually model disease processes by tracing developmental history back to its origins. Part of the Convergent Science Network podcast series from the BCBT Summer School.

Can monkeys reason logically , and if so, what does that look like at the level of single neurons? Neurophysiologist Stefano Ferraina presents evidence that prefrontal cortex neurons encode both symbolic distance and serial position during transitive inference, suggesting a neural substrate for logical reasoning in non-human primates. Subscribe for more from the Convergent Science Network podcast series. Stefano Ferraina joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss his research on transitive inference in macaque monkeys. The task requires animals to learn an ordered sequence of abstract visual symbols through pairwise comparisons, then infer the correct ranking of novel, never-before-matched pairs. Surprisingly, monkeys master this within weeks and show a robust symbolic distance effect: comparing symbols far apart in the sequence is easier and faster than comparing adjacent ones, mirroring findings in human numerical cognition. The discussion carefully examines whether this performance reflects genuine logical reasoning or simpler reward-association mechanisms. Ferraina describes a critical control experiment using two separate chains that are subsequently linked, demonstrating that monkeys maintain the transitive ordering even when reward history alone cannot explain their choices. Recording from prefrontal cortex, he finds that roughly half of task-related neurons encode the symbolic distance effect, about 40 percent encode serial position, and a subset of around 20 percent encodes both , suggesting that the same neural population supports multiple aspects of the relational structure. Key topics include how transitive inference is defined and tested in non-human primates, why the symbolic distance effect challenges pure reward-association explanations, what the serial position effect reveals about how symbols are organized along a mental continuum, how the two-chain linking experiment strengthens the case for reasoning over association, the limitations of single-neuron electrophysiology for establishing causality, and what the overlap between symbolic distance and serial position coding in prefrontal neurons implies about the neural architecture of logical inference. Part of the Convergent Science Network podcast series from the BCBT Summer School.

What happens in the brain between perfect synchrony and total disorder , and why might that intermediate zone be where cognition lives? Computer scientist Murray Shanahan explains how metastable chimera states in coupled oscillator networks may capture the dynamic coalitions that govern brain function. Subscribe for more from the Convergent Science Network podcast series. Murray Shanahan joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss his computational work on metastability and chimera states in brain-like networks. The conversation builds on Pascal Fries's communication-through-coherence hypothesis, which proposes that synchronized neuronal populations are positioned to exchange information and cooperate, while desynchronized populations are effectively shut out. Shanahan extends this framework by showing that abstract coupled oscillator models, Kuramoto oscillators, can produce chimera states where one subset of oscillators synchronizes while another remains desynchronized, and that these states are metastable, breaking apart and reforming in new configurations over time. The discussion explores how these dynamics relate to real brain phenomena, including binocular rivalry and resting-state fMRI data. When Kuramoto oscillators are placed on nodes of a real human connectome derived from diffusion tensor imaging, the model produces strong correlations with empirical resting-state functional connectivity , but only when operating in the metastable chimera regime. This finding surprised Shanahan and suggests that the brain may be poised at a critical point between order and disorder, where the richness of its dynamical repertoire is maximized. Key topics include how metastability differs from stable attractors and why it matters for cognition, what chimera states are and why physicists initially overlooked their relevance, how gamma-frequency oscillations facilitate competition and cooperation among distributed neuronal populations, why coupling strength and transmission delays are the key parameters governing these dynamics, and what the relationship is between fast oscillatory mechanisms and the slower dynamics captured by fMRI. Part of the Convergent Science Network podcast series from the BCBT Summer School.

What if the motor cortex does not just encode movements but organizes entire behavioral repertoires, reaching, grasping, defending, across three interconnected cortical stages? Neuroanatomist Jon Kaas describes how long-duration electrical stimulation reveals a modular architecture for goal-directed action in primates that challenges standard views of motor control. Subscribe for more from the Convergent Science Network podcast series. Jon Kaas joins Paul Verschure and Tony Prescott at the BCBT summer school to present his research on the functional organization of the primate motor system, spanning prosimian galagos, New World monkeys, and macaques. Using half-second electrical stimulation pulses, Kaas and colleagues discovered that specific behavioral patterns , hand-to-mouth movements, defensive gestures, reaching, grasping , can be evoked from small, corresponding regions in posterior parietal cortex, premotor cortex, and primary motor cortex. These three stages form a hierarchical but parallel system where posterior parietal cortex integrates high-level sensory information, premotor cortex contributes executive and motivational inputs, and motor cortex provides the critical output. The discussion explores how this organization differs from the standard population-vector model of motor encoding and how it relates to subcortical control. Cooling experiments demonstrate that motor cortex is required for the other stages to produce movements, confirming a hierarchical dependency. Tracer injections reveal that corresponding behavioral zones across all three cortical stages converge on the same regions of the basal ganglia, suggesting a role for subcortical structures in learning and modulating these cortical action modules. Kaas argues that posterior parietal cortex expanded dramatically in primate evolution, adding cortical control over behaviors that were previously managed subcortically. Key topics include how long-duration stimulation reveals behavioral organization invisible to standard mapping, why posterior parietal cortex is a primate innovation with multimodal sensory inputs, how inhibitory connections between behavioral zones enable competition and action selection, what the scaling challenges are from galago to human motor repertoires, and whether the modular organization of stereotyped behaviors can accommodate the arbitrary, learned action sequences that characterize human performance. Part of the Convergent Science Network podcast series from the BCBT Summer School.

What if the most widely used model of brain connectivity is too crude to capture what actually makes the cortex work? Neuroanatomist Henry Kennedy presents evidence that connection strength, not mere presence or absence of links, is where the real specificity of cortical architecture lies , spanning five orders of magnitude. Subscribe for more from the Convergent Science Network podcast series. Henry Kennedy joins Paul Verschure and Tony Prescott at the BCBT summer school to present his quantitative tract-tracing data from the macaque monkey cortex, challenging the utility of small-world network models for understanding cortical organization. With a connection density of roughly 70 percent among 91 cortical areas, Kennedy argues that binary descriptions of connectivity tell you almost nothing , at that density, everything is virtually connected to everything else. The real information lies in the weights: connection strengths that span five orders of magnitude and follow an exponential distance rule, declining sharply with the physical distance between areas. The discussion reveals that this single exponential distance rule, when used to generate random networks, reproduces many observed properties of the real cortical network , including motif distributions, clique structures, and efficiency measures under progressive thresholding. Kennedy shows that the macaque cortex achieves optimal placement of areas to minimize wiring given these weight constraints, while the mouse brain does not, suggesting fundamentally different organizational principles across species. The comparison between primate and rodent brains reveals that mice have shallower distance-decay functions, fewer cliques, and suboptimal area placement, raising serious questions about using the mouse as a model for primate cortical organization. Key topics include why weighted directed networks are more informative than binary connectivity graphs, how the exponential distance rule generates realistic cortical network properties, what optimal area placement means and how it differs between primates and rodents, why diffusion MRI cannot capture the range of connection strengths revealed by tract tracing, and how cortical folding and surface distances reshape our understanding of the distance rule across species. Part of the Convergent Science Network podcast series from the BCBT Summer School.

What if the search for a single canonical cortical microcircuit is leading neuroscience in the wrong direction? Cognitive scientist Gary Marcus argues that the brain's apparent uniformity masks functionally critical variations , and that understanding higher cognition requires computational primitives we have barely begun to identify. Subscribe for more from the Convergent Science Network podcast series. Gary Marcus joins Paul Verschure and Tony Prescott at the BCBT summer school to challenge the dominant idea that a single repeated circuit underlies all cortical computation. Drawing on evolutionary biology's principle of duplication and divergence, Marcus argues that cortical areas may share a common template but differ in ways that are functionally decisive , much like a hand and a foot share most of their genes yet serve very different purposes. He contends that the field's attraction to parsimony, while productive in physics, is misleading in biology where complexity is the rule. The discussion identifies what Marcus sees as the most critical gap in computational neuroscience: variable binding. While hierarchical feature detection is reasonably well understood and modeled, the ability to represent variables, instantiate them with particular values, maintain structured representations, and distinguish types from tokens remains unexplained at the neural level. Marcus argues these operations are non-negotiable for higher cognition, particularly language, and that no current neural network architecture adequately captures them. He also revisits his earlier claim about tree structures, now arguing that humans lack true location-addressable memory, which limits our ability to represent unbounded hierarchical structures. Key topics include why the cortex appears uniform under a magnifying glass but differs in functionally important ways, how duplication and divergence applies to cortical circuit evolution, what variable binding is and why it matters for language and reasoning, the limitations of simple recurrent networks for capturing syntax, why labeled-line architectures cannot scale to handle novel representations, and how a phylogenetic approach to cortical circuit types could advance our understanding of human cognition. Part of the Convergent Science Network podcast series from the BCBT Summer School.

What can a small monkey cracking nuts with a stone tell us about the origins of tool use , and why is there still no theory to explain it? Primatologist Dorothy Fragaszy describes how wild capuchin monkeys develop remarkably skilled percussive tool use through years of socially supported exploration, challenging assumptions about what cognition tool use requires. Subscribe for more from the Convergent Science Network podcast series. Dorothy Fragaszy joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss her fieldwork on tool use in wild bearded capuchin monkeys. These small primates, weighing only two to four kilos, routinely lift stones half their body weight to crack extremely resistant palm nuts with precision and control that takes years to develop. Fragaszy explains that tool use in non-human primates is rare , only a handful of the roughly 560 known primate species use tools habitually in the wild , and that capuchins and chimpanzees appear to have evolved this capacity independently, separated by 35 million years of divergent evolution. The conversation explores why there is no unified theory of tool use in animal behavior, and how Fragaszy draws on two theoretical frameworks: the ecological psychology of the Gibsons, emphasizing how individuals actively discover affordances in their environment, and Bernstein's work on motor coordination, addressing how a system with many degrees of freedom achieves effective, skilled action. Young capuchins are inducted into nut-cracking through a socially rich context , attracted by the sounds, sights, and smells of adult cracking activity , and progress through years of exploratory play before mastering the coordination of anvil, nut, and hammerstone. Key topics include how tool use is defined descriptively rather than theoretically in animal behavior, why extractive foraging rather than anatomical dexterity predicts which species use tools, how capuchins manage the biomechanical challenge of percussive force without injuring themselves, what the developmental trajectory from play to skilled performance reveals about perceptual learning, and why the convergent evolution of tool use across distantly related species argues against strong genetic pre-specification. Part of the Convergent Science Network podcast series from the BCBT Summer School.

Can rats imagine the future? Neuroscientist David Redish presents evidence that rodents engage in mental time travel , constructing representations of places they have not yet visited , and argues this forces us to rethink the boundaries of animal cognition. Subscribe for more from the Convergent Science Network podcast series. David Redish joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss his research on what he calls the cognitive rat. Using advanced neural decoding methods applied to hippocampal place cells, Redish demonstrates that rats generate self-consistent representations of locations they are not currently occupying , neural signatures of deliberation, imagination, and possibly insight. The conversation traces the intellectual lineage from Tolman's cognitive maps through the discovery of place cells to modern decoding techniques that allow researchers to effectively read the spatial content of ongoing neural activity. The discussion explores four distinct decision-making systems Redish identifies in the mammalian brain, reflexive, deliberative, procedural, and Pavlovian, each with largely separate neural substrates. At decision points, rats produce forward sweeps through upcoming spatial trajectories at roughly 15 times behavioral speed, while at reward locations, replay events compress spatial sequences to 40 times real time. These replay events during waking states appear to support insight and imagination, including novel shortcut sequences the animal has never physically traversed, while sleep replay tends to faithfully recapitulate actual experiences for memory consolidation. Key topics include how to define cognition operationally in non-human animals, the distinction between local and global cognitive maps, why spatial tasks reveal cognitive capacities that non-spatial paradigms miss, how mental time travel relates to episodic memory and future planning, and what the difference between waking and sleep replay tells us about the dual roles of hippocampal sharp-wave events in decision-making versus memory consolidation. Part of the Convergent Science Network podcast series from the BCBT Summer School.

How does becoming a mother permanently rewire the brain , and could the answer lie not in the genes themselves, but in how experience reshapes their expression? Neuroscientist Danielle Stolzenberg explains how epigenetic mechanisms transform the maternal brain, revealing a molecular bridge between hormones, experience, and lasting behavioral change. Subscribe for more from the Convergent Science Network podcast series. Danielle Stolzenberg joins Paul Verschure and Tony Prescott at the BCBT summer school to discuss her research on the epigenetic basis of maternal behavior in mammals. The conversation centers on a striking biological puzzle: most mammalian females undergo a dramatic shift in responsiveness around the time of birth, and that change persists for life. Stolzenberg investigates how brief experiences with infants, combined with hormonal priming, produce long-lasting changes in gene expression through histone acetylation , a form of cellular memory that alters how neurons in the maternal circuit function without changing the DNA sequence itself. The discussion unpacks how estradiol and oxytocin prepare the brain for motherhood, but experience with pups is what consolidates the behavioral transformation. Stolzenberg presents evidence using histone deacetylase inhibitors to show that enhancing histone acetylation can accelerate maternal learning, reducing the number of pup exposures needed to produce lasting maternal responsiveness. Her work targets the medial preoptic area of the hypothalamus, a region central to maternal care, where she has identified increases in CREB-binding protein following infant interaction. Key topics include the molecular distinction between genetics and epigenetics, how hormones and experience converge on shared chromatin-modifying pathways, the role of CREB-mediated gene transcription in memory consolidation, and why maternal behavior serves as a powerful model for understanding how transient experiences produce permanent changes in brain function. The conversation also addresses whether these epigenetic modifications could be transmitted across generations and what that means for understanding behavioral inheritance. Part of the Convergent Science Network podcast series from the BCBT Summer School.

How does a single Purkinje cell receiving half a million inputs learn to produce a precisely timed eye blink, and why has the cerebellum been so difficult to understand despite its crystalline simplicity? Germund Hesslow reveals what decades of painstaking physiology have uncovered.Subscribe for more from the Convergent Science Network podcast series.Germund Hesslow describes his journey from Freudian psychology to hardcore cerebellar physiology, drawn by the exceptional quality of research in the Lund laboratory rather than the subject matter itself. The serendipitous discovery that Pavlovian conditioning occurs in the cerebellum united his interests in learning with his technical expertise, launching a 25-year investigation into how associative memory is formed at the cellular level. Working with a decerebrate preparation that retains only the cerebellum and brainstem, Hesslow's group demonstrates that conditioning follows essentially the same rules as in intact animals, taking similar time to acquire and extinguish.The episode provides a clear account of the cerebellar circuit for eye-blink conditioning. Conditioned stimulus information arrives via mossy fibers and parallel fibers, converging on Purkinje cells that also receive climbing fiber input carrying the unconditioned stimulus signal. With each Purkinje cell receiving up to half a million parallel fiber inputs carrying information about virtually everything happening to the organism, the system is ideally suited for forming associations. The learned response manifests as a precisely timed pause in the Purkinje cell's tonic inhibitory output, which disinhibits the deep cerebellar nuclei to generate the conditioned eye blink.The central unsolved problem is timing. The conditioned response is adaptively timed to the interstimulus interval: train with 300 milliseconds and the response peaks at 300 milliseconds; train with 500 milliseconds and it shifts accordingly. Hesslow's recent experiments systematically eliminate candidate timing mechanisms. Stimulating mossy fibers directly still produces timed responses, ruling out delays in the input pathway. Stimulating parallel fibers directly yields the same result, eliminating delays in the granule cell layer. The timing mechanism must reside close to the Purkinje cell itself, either in cortical interneurons or in the cell's intrinsic properties.Hesslow also raises the provocative possibility that different zones of the cerebellum, despite their apparently uniform crystalline structure, may operate with different temporal and functional properties, potentially explaining decades of disagreement between researchers working on eye-blink conditioning, vestibulo-ocular reflex adaptation, and in vitro slice preparations in different cerebellar regions.

Why does the prefrontal cortex prefer not to be involved, and how does a cascade of cognitive control from premotor cortex to frontal pole organize human decision-making? Etienne Koechlin maps the hierarchical architecture of executive function.Subscribe for more from the Convergent Science Network podcast series.Etienne Koechlin presents a model of prefrontal cortex function centered on the idea that action imposes a fundamental constraint on cognition: it forces the brain to collapse multiple interpretations into a single committed choice. Rather than viewing the prefrontal cortex as a repository of complex representations, Koechlin argues its primary role is to introduce seriality and decisiveness into cognitive processing, excluding alternative interpretations so the organism can act. The system's default state is automated behavior driven by premotor and posterior associative regions; the prefrontal cortex engages only when these routines fail.The hierarchical organization follows a posterior-to-anterior gradient with three distinct levels. The premotor cortex stores basic stimulus-response associations. When these are ambiguous, the posterior prefrontal cortex incorporates immediate contextual cues from the present environment. When context is insufficient, more anterior regions access episodic information from the past. At the apex, the frontal pole enables the consideration of multiple alternative strategies simultaneously, breaking the pure seriality that characterizes lower levels. Koechlin emphasizes that these levels operate concurrently rather than sequentially, with the system recruiting more anterior regions only as needed.A key function Koechlin attributes to the prefrontal cortex is monitoring, specifically judging whether a current behavioral strategy remains reliable based on its ability to predict action outcomes. He distinguishes between relative monitoring, which compares alternatives against each other, and absolute monitoring, which evaluates each strategy independently against a reliability criterion. The absolute approach avoids the trap of being locked into a limited set of alternatives and enables the critical decision of whether to persevere with learning or abandon a strategy entirely.The episode reveals an intriguing limitation: humans can suspend one task to perform a subtask, but attempting a second level of recursive suspension produces severe performance deficits, suggesting the monitoring system operates at only one level without true recursion. Koechlin connects this to the broader question of how automatization transfers complex behaviors from prefrontal control to encapsulated routines in premotor and posterior cortical regions.

Can a monkey learn to control a single neuron in its motor cortex independently of the muscles it normally drives? Eberhard Fetz traces five decades of work on volitional neural control from biofeedback to brain-computer interfaces.Subscribe for more from the Convergent Science Network podcast series.Eberhard Fetz recounts the intellectual journey from his pioneering 1969 experiments on operant conditioning of single motor cortex neurons to the development of the Neurochip, an autonomous neural interface that creates artificial connections between brain sites, spinal cord, and muscles. His early work demonstrated that monkeys could learn to volitionally increase or decrease the firing rate of individual neurons to earn rewards, with the key insight that the animal was not simply conditioning a neuron in isolation but learning to control the pattern of activation flowing through a fixed circuit in novel ways.A central theme is the remarkable flexibility of neural control. Fetz showed that neurons with consistent relationships to specific muscles could be operantly dissociated: a cell that always co-activated with the biceps could be driven to fire without any muscle activity, and vice versa. He interprets this not as rewiring of anatomical connections but as the brain exploiting its existing circuitry in variable patterns, analogous to trains taking different routes over fixed railroad tracks. This distinction between structure and activation patterns has profound implications for understanding how the brain achieves flexible behavior without constantly modifying its physical connectivity.The Neurochip technology, developed with Andy Jackson and Jaideep Mavoori, enabled a breakthrough: autonomous, battery-powered devices mounted on the monkey's skull could record neural activity and deliver spike-triggered stimulation during days of free behavior. This allowed the creation of artificial recurrent connections between cortical sites, between cortex and spinal cord, and between cortex and muscles. In one paradigm, monkeys with temporarily paralyzed wrist muscles learned to drive a cursor into targets by activating motor cortex cells that directly stimulated the denervated muscles through the Neurochip.The episode explores the relationship between volitional control and mental imagery, the challenges of reverse recruitment order when electrically stimulating muscles, and the potential for bidirectional brain-computer interfaces to restore function after spinal cord injury. Fetz draws a provocative parallel between the self's relationship to the brain and the brain's relationship to external devices, suggesting that the same mechanisms of flexible volitional control apply in both domains.

Why can you smell a molecule you have never encountered before, and how does the nose use antagonism, binding proteins, and chemotopic maps to decode the chemical world? Tim Pearce explores the engineering principles behind biological olfaction.Subscribe for more from the Convergent Science Network podcast series.Tim Pearce provides a comprehensive tour of natural olfaction, from the molecular interactions at the receptor sheet to the computational principles that enable detection of thousands of diverse chemical compounds. He highlights the system's remarkable foreignness property: unlike vision with its handful of receptor types, olfaction deploys over one percent of the genome to create a broad, relatively unbiased sampling of chemical space, capable of responding to novel molecules never previously encountered by the species.The episode reveals several layers of molecular complexity that precede neural processing. Odorant binding proteins in the nasal mucosa act as selective transporters, shifting sorption spectra to capture hydrophobic compounds that would otherwise resist the liquid phase. Odor degrading enzymes terminate signals in a timely fashion. Most surprisingly, recent evidence shows that receptor-ligand interactions involve not just affinity but also efficacy, and that widespread antagonism between molecules means the neural response to mixtures is far from a linear sum of individual components. These nonlinear competitive interactions at the receptor level fundamentally shape the olfactory code.Pearce describes a chemotopic organization of the receptor sheet where molecular features like carbon chain length, functional groups, and hydrophobicity map onto different spatial zones, driven partly by differential sorption along the airflow path and partly by the zonal expression of receptor families including ancient fish-derived class 1 receptors. Analysis of molecular descriptors reveals that despite hundreds of possible chemical features, the effective dimensionality of odor space is surprisingly low, with principal components capturing much of the perceptual variance.The discussion also covers retronasal olfaction, where volatile compounds from food reach the receptor sheet through the back of the nose and produce qualitatively different percepts than the same compounds delivered orthonasally, even at the level of receptor sheet activation patterns. Pearce connects these biological insights to engineering principles for artificial olfactory systems, arguing that the conserved architectural motifs found across species from insects to mammals provide a blueprint for building chemical sensing systems.

What if planning is not about computing value functions but about performing probabilistic inference? Marc Toussaint shows how recasting optimal control as message passing opens new computational pathways for robotics and decision-making.Subscribe for more from the Convergent Science Network podcast series.Marc Toussaint presents a theoretical framework that reformulates planning and optimal control as probabilistic inference in graphical models. Rather than iterating backward through Bellman equations to compute value functions, his approach computes both forward and backward messages whose product yields a posterior distribution over actions. This shift in perspective is not merely notational: it leads to genuinely different approximation algorithms, particularly for complex problems like partially observable Markov decision processes and factored state spaces where traditional value function methods struggle.The conversation traces the intellectual lineage from Kalman's duality between control and filtering through Bert Kappen's work on path integrals to Toussaint's own generalization that operates over joint state-control processes without restrictive assumptions about dynamics or cost structure. A key theoretical achievement is demonstrating that many existing reinforcement learning algorithms emerge as special cases of this unified formulation, providing both theoretical elegance and inherited empirical validation.Toussaint derives a model-free reinforcement learning algorithm from this framework where the policy is represented as a Boltzmann distribution. Analysis of its fixed-point properties reveals a surprising result: for non-optimal actions, the Boltzmann energy diverges to negative infinity, making them vanishingly improbable, while for optimal actions, it converges exactly to the optimal value function. The framework handles goal conflicts through the natural machinery of probabilistic inference, where inconsistent evidence simply reduces likelihood and the system finds probabilistic compromises.The episode also explores Toussaint's robotics applications, where model-based approaches using stochastic relational rules enable robots to generalize from minimal experience. Active exploration strategies that maximize information gain prove essential in the exponentially large state spaces created by relational representations of multi-object environments, allowing a robot that has observed balls rolling to intelligently seek out non-ball-shaped objects to test next.

How do the hippocampus and ventral striatum coordinate to tag locations with reward value, and what happens to place cells when something motivationally significant changes? Cyriel Pennartz reveals population-level state transitions in the rat brain.Subscribe for more from the Convergent Science Network podcast series.Cyriel Pennartz presents a detailed picture of how the rat brain's cognitive architecture processes spatial, motivational, and action-related information through the cortico-basal ganglia-hippocampal loops. He describes a continuous topographic organization where the dorsolateral striatum handles detailed sensorimotor associations and habits, the ventromedial striatum processes action-outcome relationships, and the ventral striatum integrates spatial and motivational cues. Rather than supporting a strict actor-critic division, Pennartz argues for more homogeneous computational principles operating across the striatum, with different loops processing different content but using similar mechanisms.A central finding concerns how hippocampal place cells and ventral striatal neurons respond to motivationally relevant events. Recording from approximately 600 neurons simultaneously, Pennartz discovered that reward-predictive cue lights trigger coordinated state transitions across both structures. Using K-means clustering in high-dimensional neural state space, he identified moments where the population activity undergoes a coherent shift, with a majority of cells showing marked firing rate changes. These transitions occur not only in response to explicit cues but also spontaneously when the rat enters reward-associated chambers, and they are correlated between hippocampus and ventral striatum.The episode explores an intriguing observation about hippocampal place field properties: reward sites attract unusually small, precise micro-place fields compared to the larger fields found in non-rewarded compartments. Pennartz suggests this finer spatial scaling reflects both the behavioral complexity at reward sites and the biological importance of precise spatial knowledge at these locations. He proposes that the hippocampus provides a spatio-temporal scaffold onto which motivationally significant events are tagged, analogous to the ancient Roman memory palace technique.The discussion also addresses the four key information domains processed through these loops: cues, actions, motivation, and space, with time emerging as a potential fifth dimension handled through ramping firing rate responses and possibly cerebellar timing circuits operating at finer temporal resolutions.

What if consciousness evolved not to perceive the world but to survive in a world full of other minds? Paul Verschure proposes that the unified conscious scene solves a credit assignment problem created by parallel social simulations.Subscribe for more from the Convergent Science Network podcast series.In this episode, Paul Verschure is interviewed by Tony Prescott and Tim Pearce about his theory of consciousness and its relationship to his Distributed Adaptive Control (DAC) architecture. Verschure begins by surveying the landscape of consciousness research, identifying five families of necessary but insufficient conditions: embodied grounding (Metzinger, Damasio), sensorimotor coupling (O'Regan), predictive simulation (Hesslow), integration and differentiation (Tononi, Edelman), and global workspace dynamics (Baars, Dehaene). He argues that each captures a real feature of conscious processing but none alone is sufficient.The DAC architecture provides the broader framework: a layered control system with reactive, adaptive, and contextual layers, crossed by columns processing world states, self states, and action. Verschure argues this architecture handles the H4W problem of interacting with the physical world (why, what, where, when) but does not require consciousness. The critical transition occurs during the Cambrian explosion when organisms suddenly faced a world populated by other agents whose internal states, goals, and strategies are hidden from surface observation.Verschure's central hypothesis is that consciousness evolved to solve the credit assignment problem created by running multiple parallel simulations of other agents' intentions. Real-time behavior is controlled by these parallel loops, but their outputs may conflict. The unified conscious scene serves as a delayed but coherent compression of all simulations into a singular assessment of what is actually happening, collapsing the possible into the actual. This singular state can then propagate value signals back to the parallel controllers, optimizing their future performance. The conscious scene runs behind real time, consistent with Libet's findings, but serves a genuine causal function rather than being epiphenomenal.The episode includes a critical examination of Tononi's integrated information theory, where Verschure argues that phi-like measures of neural variability fail to distinguish between pre-conscious states with multiple competing options and the unitary conscious scene that emerges after competitive selection.

Can a crow that has never seen a particular problem still build the right tool to solve it, and what does that tell us about the nature of animal intelligence? Alex Kacelnik explores the boundaries between insight and learning in New Caledonian crows.Subscribe for more from the Convergent Science Network podcast series.Alex Kacelnik brings a biologist's perspective to animal cognition, positioning intelligence as an evolved toolkit shaped by natural selection rather than an abstract capacity to be ranked on a human-centric scale. He draws a critical distinction between risk, where probabilities are known, and uncertainty, where even the nature of the problem is unclear, arguing that learning transforms individual uncertainty into manageable risk by filling in the parameters that evolution could not anticipate.The centerpiece of the discussion is the New Caledonian crow, the most intensely tool-dependent non-human species known. These birds manufacture at least five categories of tools including hooks, straight sticks, and elaborately shaped pandanus leaf strips, showing regional variation that suggests cultural transmission. In laboratory settings, the crows demonstrate remarkable flexibility: they select tools of appropriate length and diameter for specific problems, build hooks when straight tools will not work, and solve multi-step problems requiring sequential tool use on a trial-unique basis. Kacelnik emphasizes that these behaviors cannot be fully explained by chaining previously reinforced responses, as the complete sequences have never been experienced before.Yet Kacelnik resists easy mentalistic interpretations. He positions himself closer to the "killjoy behaviorist" than the "mystical psychologist," insisting that terms like insight, planning, and understanding should only be used when backed by algorithmic models of how experience translates into novel solutions. A key experiment illustrates this principled caution: crows could innovate by dropping stones into a mechanism to release food, but only if they had prior experience with how the magnetic release mechanism worked. Innovation requires partial knowledge as scaffolding, not magical leaps of comprehension.The episode also examines how crows use tools not just for food extraction but for exploring potentially dangerous objects at a safe distance, and how sexual selection in siskins illustrates the complex evolutionary pressures shaping cognitive abilities across bird species.

Why does the monkey prefrontal cortex keep future goals and past goals in separate neural populations, and what does the frontal pole exclusively care about? Aldo Genovesio reveals how the primate brain organizes goal-directed behavior.Subscribe for more from the Convergent Science Network podcast series.Aldo Genovesio presents neurophysiological findings from single-cell recordings in the monkey prefrontal cortex that illuminate how the brain represents goals, strategies, and task monitoring. Using a strategy task where monkeys must remember previous goals to determine future actions, his laboratory discovered that prefrontal neurons encode conjunctions of abstract information: individual cells combine representations of strategy (repeat-stay or change-shift) with specific goals or stimulus features, revealing a rich combinatorial code for task-relevant variables.A striking organizational principle emerges from the data: neurons encoding future goals and neurons encoding past goals form separate, non-overlapping populations within the same prefrontal region. Future goal cells show correlated activity with each other, suggesting they form a coherent network capable of driving premotor cortex toward action selection. Past goal cells, by contrast, show no such inter-neuronal correlation. Genovesio interprets this segregation as potentially facilitating output monitoring, the ability to distinguish accomplished goals from pending ones, a function known to be impaired in patients with prefrontal damage and dementia.The conversation takes a surprising turn with findings from the frontal pole, the most anterior region of the cortex. Recording from hundreds of neurons, Genovesio found that approximately 30 percent encode a pure monitoring signal: they respond exclusively during feedback about whether the monkey succeeded or failed, with no representation of stimuli, strategies, future goals, or past goals. This extreme selectivity contrasts sharply with the mixed representations found in more posterior prefrontal regions and suggests that the frontal pole performs a highly specialized abstraction rather than simply integrating more information as hierarchical models might predict.The episode raises fundamental questions about how the brain transitions goal representations from future to past status, whether the frontal pole's monitoring signal serves as a gate for updating goal networks, and how these findings relate to the broader hierarchical organization of the frontal lobe.

How does a mouse nose with 1,200 receptor genes wire itself into a precise sensory map, and why is that map less stereotyped than we once believed? Peter Mombaerts explores the genetics and development of olfactory circuit formation.Subscribe for more from the Convergent Science Network podcast series.Peter Mombaerts describes the remarkable complexity of the mouse olfactory system, where approximately 1,200 odorant receptor genes each expressed by a distinct population of sensory neurons must organize their axonal projections into roughly 3,600 glomeruli in the olfactory bulb. Using gene targeting and molecular labeling techniques, his laboratory has spent two decades investigating how this glomerular map develops and what role the receptor proteins themselves play in axon guidance and glomerular identity.A central theme is the surprising degree of variability in glomerular positioning. Mombaerts challenges the widely used term "stereotyped" to describe the glomerular map, demonstrating that even between the left and right bulbs of the same inbred mouse, the relative positions of identified glomeruli can be inverted. He prefers terms like "recognizable" or "reproducible," noting that the precision is insufficient to construct a definitive atlas as has been done for Drosophila. This variability has important implications for understanding the mechanisms of map formation: if the map were truly stereotyped, extremely complex molecular guidance mechanisms would be required, but acknowledging the jitter relaxes these demands considerably.The conversation explores the genomic organization of odorant receptor genes, which Mombaerts describes as "haphazard," distributed across approximately 40 loci with the largest cluster containing around 300 genes. Expression levels vary over two orders of magnitude between different receptor types, and the one-neuron-one-receptor rule, while strongly supported, remains an asymptotic conclusion. Remarkably, replacing an odorant receptor's coding region with the beta-2 adrenergic receptor still produces neurons that form a recognizable glomerulus and respond to appropriate ligands, suggesting that the receptor protein's role in axon guidance may not be unique to olfactory receptors.The episode also addresses the emerging recognition that odorant receptor genes are expressed outside the nose, including in kidneys where knockout of one receptor affects blood pressure regulation, hinting at broader biological roles for this gene family beyond olfaction.

How does a 560-million-year-old fish illuminate the control architecture behind all vertebrate movement? Sten Grillner traces the neural circuits of locomotion from lamprey spinal cord to human basal ganglia.Subscribe for more from the Convergent Science Network podcast series.Sten Grillner presents decades of work on the lamprey, a jawless fish that emerged during the Cambrian explosion, as a model for understanding the conserved control systems underlying vertebrate motor behavior. He explains how the lamprey's spinal cord contains approximately 100 central pattern generators (CPGs) that produce rhythmic swimming through the interplay of excitatory premotor interneurons, inhibitory coordination neurons, and critical membrane properties including NMDA receptors, voltage-dependent calcium channels, and calcium-activated potassium channels. Even without sensory feedback, the isolated spinal cord generates well-coordinated locomotor patterns, though stretch receptors provide essential compensation for environmental perturbations.The conversation reveals how detailed computational models of the lamprey spinal cord, incorporating biological variability in cellular properties across neuron populations, demonstrated that this variability is not noise but a design feature essential for stable motor output. A striking finding from large-scale simulations with 10,000 neurons showed that modifying just 5-10 percent of the network could entirely transform the pattern of activity, enabling transitions between forward and backward swimming.Grillner then ascends the neural hierarchy to describe how basal ganglia circuits control behavior through a layered architecture. The substantia nigra reticulata and globus pallidus project directly to brainstem locomotor command centers and the tectum, providing powerful inhibitory gating of motor programs. The striatum receives cortical input and interfaces with the thalamus in recurrent loops. He presents this as a four-layered control structure: CPGs at the base, brainstem motor nuclei, the nigra-pallidus output layer, and the cortex-striatum input layer, with the thalamus providing modulatory feedback across the upper layers.The discussion explores how this basic architecture has been elaborated through vertebrate evolution, from the emergence of paired fins in elasmobranchs to the development of limbs in tetrapods, while the fundamental circuit principles remain remarkably conserved. Grillner argues that new motor capabilities arise not from qualitative changes in spinal circuitry but from the parceling out of interneuron populations to independently control new appendages.

Can the way we perceive forces explain how we understand both physical actions and social interactions? Peter Gardenfors extends his conceptual spaces framework from static objects to the dynamic world of action and events.Subscribe for more from the Convergent Science Network podcast series.Peter Gardenfors introduces his theory of conceptual spaces as a geometric approach to knowledge representation that sits between symbolic AI and neural networks. The framework organizes knowledge along quality dimensions grouped into domains, such as the three-dimensional color space of hue, brightness, and saturation. Concepts correspond to convex regions in these high-dimensional spaces, with prototypes at their centers of gravity. Gardenfors describes the framework as neo-Kantian: some domains are innate, tied to our sensory organs, while others are culturally acquired and can expand throughout development.The conversation takes a fascinating turn when Gardenfors extends this framework to action representation. Rather than treating actions as static classifications, he proposes that we perceive actions primarily through patterns of force, specifically the second derivative of movement. Drawing on Gunnar Johansson's point-light display experiments showing that humans identify biological motion from minimal cues within 200 milliseconds, Gardenfors argues that our brains impose a notion of force as a kind of dynamic contour on perceived movement. This force-based interpretation extends beyond Newtonian physics to encompass social forces like authority and attraction, suggesting that the brain assigns pseudo-causal relationships to observed changes regardless of their true physical origins.Gardenfors then develops a minimal theory of events built on two vectors acting on a patient: a force vector describing what causes change and a result vector describing the change itself. This decomposition handles cases from simple physical interactions to events where forces balance and nothing happens, though Gardenfors acknowledges that highly abstract events like the Olympics remain challenging for the framework. The discussion explores how this event structure maps onto language, with the patient, agent, force, and result components providing a cognitive foundation for how we construct and understand sentences about the world.Throughout the episode, the interplay between perception, action, and language emerges as a central theme, with conceptual spaces serving as a modality-independent representational engine that bridges bottom-up sensory processing and top-down symbolic reasoning.

Where in the brain does the meaning of a word live, and why does hearing "kick" activate your leg motor cortex? Friedemann Pulvermuller unpacks how the brain grounds language in sensory and motor experience.Subscribe for more from the Convergent Science Network podcast series.Friedemann Pulvermuller presents a neurobiological account of word meaning that challenges traditional modular theories of semantics. Drawing on his mentor Valentino Braitenberg's vision of the cortex as an information mixing system, Pulvermuller argues that meaning arises from distributed cortical circuits where neurons that were originally specialized for vision or motor control become cross-modal through mutual linkage. The result is that understanding a word like "grasp" activates hand motor representations, while "kick" engages leg-related cortical areas, with activation patterns overlapping those produced by actual movements.The conversation carefully distinguishes four facets of semantics: referential, combinatorial, abstract, and emotional. Referential semantics connects words to objects and actions in the world, solving the symbol grounding problem that purely symbolic approaches cannot address. Combinatorial semantics captures statistical co-occurrence patterns between words, allowing even a blind person to learn that strawberries are red. Abstract semantics, illustrated through the concept of freedom, requires more computational power because multiple diverse prototypical instantiations must be linked through logical either-or operations. Pulvermuller acknowledges these categories represent extremes on a continuum rather than hard boundaries.The empirical evidence builds from early EEG studies showing differential hemispheric activation for content versus function words, through PET studies of tool and animal naming, to the critical finding that action verbs related to different body parts produce somatotopically organized activation in motor cortex. This body-part specificity, controlled for linguistic confounds like imageability and grammatical class, provided the strongest evidence that semantic processing engages sensorimotor systems in a content-specific manner.Pulvermuller frames his approach within a Braitenberg-inspired correlation learning framework, where Hebbian strengthening of connections between co-active neural populations creates the distributed circuits that carry meaning, offering a mechanistic bridge between neural anatomy and the richness of human language.

What can a 535-million-year-old fossilized brain tell us about the origins of our own nervous system? Nick Strausfeld reveals how ancient arthropod fossils are rewriting the evolutionary history of the brain.Subscribe for more from the Convergent Science Network podcast series.Nick Strausfeld makes a compelling case for why neuroscience must be grounded in evolutionary and comparative biology. He argues against the over-reliance on a handful of model organisms, insisting that understanding the brain's design principles requires studying nervous systems across a wide range of species. The conversation traces the deep architectural features shared by insect and crustacean brains, revealing a common organizational template built around glomerular processing units that can serve olfactory, visual, or tactile modalities with fundamentally similar computational circuits.Strausfeld describes a hierarchical brain architecture where sensory-specific processing occurs at peripheral levels while higher centers like mushroom bodies and the central body complex provide substrates for allocentric memory, behavioral choice, and complex decision-making. He proposes that these integrative structures may derive from an ancient, pre-segmental ancestor shared with polychaete worms, representing a "brain within the brain" that was later incorporated into the arthropod head. The discussion explores how ecological pressures drive variation in neural organization, with examples of how different fly species show divergent lobular plate architectures corresponding to their distinct flight behaviors.The most striking revelation concerns Strausfeld's discovery of fossilized brains from the Cambrian period. Working with specimens from the Chengjiang mudstone dating to 535 million years ago, he identified a stem-group arthropod called Fuxianhuia whose brain shows three fused ganglia and three nested optic neuropils characteristic of modern crustaceans, despite having an extremely simple body plan. This finding overturns the assumption that branchiopods represent the ancestral condition and demonstrates that sophisticated neural architecture preceded the explosive diversification of body forms during the Cambrian.The episode challenges the intuition that brains evolve from simple to complex, highlighting examples of evolutionary reversal and loss, and argues that direct anatomical evidence from fossils is essential for reconstructing neural evolution in ways that molecular phylogenetics alone cannot achieve.

How does your brain decide what you're seeing before you've even finished looking? Moshe Bar reveals how the orbital frontal cortex uses blurry, low-resolution snapshots of the world to generate rapid predictions that shape perception in real time.Subscribe for more from the Convergent Science Network podcast series.In this episode, Moshe Bar challenges the textbook separation between perception and cognition, arguing that these processes are deeply intertwined rather than sequential. He presents evidence that the orbital frontal cortex (OFC) receives coarse, low spatial frequency visual information and uses it to generate top-down predictions that actively guide how we perceive our environment. Bar estimates the balance between bottom-up sensory input and top-down prediction can range from zero to one hundred percent depending on context, from meditative states where expectations are silenced to planning scenarios driven entirely by internal models.Bar describes how faces can be categorized as threatening or non-threatening in as little as 39 milliseconds using low spatial frequency information, with the amygdala playing a key role. He positions the OFC not as a purely visual area but as a polysensory prediction hub that integrates subcortical and cortical inputs to anticipate what is coming next across multiple timescales. The discussion explores how OFC predictions relate to contextual memory networks involving medial prefrontal cortex, parahippocampal cortex, and retrosplenial cortex, each contributing different aspects of scene understanding from abstract schemas to specific spatial details.A particularly compelling segment examines how contextual associations are organized in the brain. Using MEG phase-locking analysis and Granger causality, Bar shows that highly contextual objects activate a tightly synchronized three-node network, while non-contextual objects do not produce the same coherent activation. The conversation also addresses how spatial and temporal dimensions of context are processed, and how ambiguous stimuli like the word "bank" require the brain to activate and then suppress competing context frames.Bar's work raises fundamental questions about the evolutionary origins of rapid prediction, the relationship between the OFC and amygdala as parallel threat-assessment systems, and whether the brain's predictive machinery extends beyond vision to prepare the body for action across all sensory modalities.

A honeybee learns a color in five visits, generalizes matching rules across sensory modalities, and signals food distance to nestmates through dance. How does a brain with fewer than a million neurons achieve cognitive feats that challenge our understanding of intelligence?Subscribe for more from the Convergent Science Network podcast series.Srinivasan explains that insect compound eyes create a fundamentally different visual world than vertebrate camera eyes. With the two compound eyes too close together for effective stereo vision, bees rely on optic flow, the apparent motion of images across the retina during flight, to gauge distance. His tunnel experiments demonstrated that bees measure distance in units of integrated optic flow rather than absolute meters, which means flying over a featureless lake versus a textured forest produces different distance readings. The system works because all bees from the same hive take the same route, so calibration errors cancel out in the waggle dance communication.The waggle dance itself encodes both direction and distance to food sources. Direction is referenced to the sun's position or the sky's polarization pattern, while distance is conveyed by the duration of the waggle run. Srinivasan describes how recruited bees evaluate the ratio of caloric return to energy expenditure, effectively performing cost-benefit analysis before choosing which advertised food source to visit. Intriguingly, angular precision in the dance increases with distance, compensating for the fact that a fixed angular error maps to a larger search area at greater range. The evolutionary origins of the dance may trace to solitary butterflies that perform waggle movements without an audience, suggesting the behavior was co-opted for communication from a pre-existing motor pattern.The cognitive capabilities of bees extend far beyond navigation. They learn colors in five rewards, discriminate wavelengths with near-human precision, and exhibit color constancy across lighting conditions. Most remarkably, bees trained on a delayed match-to-sample task using odors spontaneously transfer the matching rule to visual stimuli they have never been trained on, demonstrating cross-modal concept learning. The mushroom bodies, which expand dramatically when bees begin foraging, likely serve as the invertebrate analog of the hippocampus, though the physiological basis of bee memory remains almost entirely unknown.

How do two tango dancers achieve millisecond-level coordination without a conductor? Guenther Knoblich decomposes joint action into five mechanisms, from unconscious entrainment to motor simulation, revealing that even speeding up is a sophisticated coordination strategy.Subscribe for more from the Convergent Science Network podcast series.Knoblich defines joint action broadly as any coordination between people in space and time, deliberately avoiding distinctions between intentional and unintentional, cooperative and competitive. This breadth allows him to identify shared mechanisms across seemingly different situations: table tennis opponents and dance partners may rely on the same low-level coordination processes despite having opposing goals. He identifies five mechanisms ranging from simple to cognitively demanding: entrainment, speeding, simulation, monitoring, and signaling.Entrainment, borrowed from physics, describes how oscillating systems with perceptual coupling tend to synchronize automatically. People walking near each other converge on the same pace without intending to; rocking chairs in the same room align their rhythms. But Knoblich argues entrainment alone cannot explain most joint action. His group discovered speeding as an independent strategy: when asked to synchronize discrete responses with a partner, people speed up by about 50 milliseconds compared to individual performance. This is not competition or arousal. Correlation analysis reveals a causal chain: faster reactions reduce variability, and reduced variability decreases asynchrony between partners. The effect appears immediately and remains constant, suggesting a general mindset shift rather than a learned adjustment.The discussion of motor simulation draws on EEG evidence from a bottle-passing task. The receiver shows motor preparation peaks time-locked to the giver's action initiation, well before their own receiving movement begins, demonstrating that the motor system predicts a partner's actions in parallel with planning one's own. Knoblich proposes that the same forward models used for individual action planning are repurposed to simulate others, with expertise modulating simulation fidelity: an expert dancer simulates observed dance movements with greater motor activation than a novice. This framework connects individual motor control to social cognition through shared predictive mechanisms rather than requiring a separate theory-of-mind module.

Watch an expert rock climber study a wall they have never seen before, and you are watching the motor system think. Giovanni Pezzulo explains how the predictive brain reuses sensorimotor knowledge for problem solving, imagery, and understanding other minds.Subscribe for more from the Convergent Science Network podcast series.Pezzulo distinguishes two kinds of prediction that are often conflated in the literature. Implicit prediction, as in classical conditioning, attaches value labels to stimuli without maintaining an internal model of the predictive relationship. Explicit prediction builds structured forward models of environmental regularities that can be run offline for planning, decision-making, and mental simulation. The predictive brain hypothesis proposes that the brain systematically incorporates environmental structure into such models and uses them to drive perception, attention, and action selection proactively rather than reactively.The interview centers on embodied problem solving, illustrated by competitive rock climbers who study an unfamiliar wall before ascending. Expert climbers visibly rehearse motor sequences, moving their arms to simulate grasps and reaches, using their bodies as external scaffolds for cognition. This is not mere motor programming: the climber must assemble partial skills in novel combinations, evaluate reachability constraints, and explore a vast space of possible routes, all guided by proprioceptive knowledge that only expertise provides. Memory experiments confirm that expert climbers remember difficult routes significantly better than novices, but only when the routes are actually climbable, demonstrating that motor expertise structures perception and memory rather than providing a generic cognitive advantage.Pezzulo builds from individual action to social cognition through a series of escalating steps. If your motor system generates predictions about your own actions, it can also predict the actions of others by running the same forward models with different parameters. This simulation-based understanding of others bootstraps joint action planning, coordination, and eventually the ability to influence another person's beliefs and intentions. Clinical evidence supports this continuum: patients with bilateral parietal lesions cannot inhibit imagined actions from becoming overt movements, and utilization behavior patients automatically grasp objects they see, revealing the tight coupling between internal simulation and motor execution that normally remains covert.

Beneath every thought, every emotion, and every decision lies a primitive engine that neuroscience has ignored for 60 years. Donald Pfaff makes the case that generalized arousal is the essential foundation of all brain function, from fear to physics exams.Subscribe for more from the Convergent Science Network podcast series.Pfaff argues that arousal has been wrongly dismissed as non-specific background noise. He reframes it as the necessary precondition for all motivated behavior: alertness to sensory stimuli, motor activity, and emotional reactivity. The hyperthyroid individual who responds to every stimulus, cannot stand still, and weeps or laughs readily exemplifies high arousal; the hypothyroid couch potato who is sluggish, unreactive, and emotionally flat exemplifies the opposite. His high-throughput behavioral assay measures mice in isolation across sensory responsiveness, locomotion, and conditioned fear responses, 50 times per second, 24 hours a day, seven days a week.Covariance analysis across multiple arousal-related tests reveals that generalized arousal accounts for approximately 30% of behavioral variance, a substantial foundation upon which specific drives like hunger, fear, and sex layer additional motivation. Pfaff frames this quantitatively: for the act of raiding the refrigerator at midnight, generalized arousal contributes roughly 30%, hunger drive perhaps 50%, personality factors another portion, with an irreducible margin of error that should trouble any judge deciding capital punishment cases.The neural substrate involves both ascending and descending pathways. Five ascending neuromodulatory systems, norepinephrine, dopamine, serotonin, histamine, and acetylcholine, project from brainstem to forebrain, each clinically familiar through drugs that manipulate them. Pfaff distinguishes a phylogenetically ancient low road through the hypothalamus and basal forebrain from a high road through the thalamus to cortex. Giant neurons in the nucleus gigantocellularis of the brainstem reticular formation may serve as critical hubs, projecting both rostrally and caudally, linking arousal to both cortical activation and autonomic control. The neuropeptide CRF, operating through three receptor types, provides a specific neuromodulatory mechanism for danger-related arousal.

Can the brain's visual wiring be explained by the same engineering principles that optimize telephone networks? Dmitri Chklovskii shows how predictive coding theory and lattice filters map onto real neural circuits, from fly photoreceptors to the mammalian LGN.Subscribe for more from the Convergent Science Network podcast series.Chklovskii bridges theoretical physics and neuroscience by applying adaptive signal processing frameworks to sensory systems. Building on Barlow's redundancy reduction principle and the predictive coding work of Srinivasan, Laughlin, and Dubs, his group derives normative predictions for neural filter shapes with no free parameters: once you specify the natural stimulus statistics and signal-to-noise ratio, the optimal filter is uniquely determined. The biphasic temporal response and center-surround spatial receptive fields of retinal and LGN neurons emerge naturally as mechanisms for subtracting predictions from incoming signals, compressing redundant information.The key evidence supporting this framework over simple biophysical explanations like after-hyperpolarization comes from stimulus-dependent filter changes. At high contrast, neurons show sharp biphasic responses with strong negative components; at low contrast, the filter shifts toward broader low-pass characteristics with weakened negative phases. This adaptive behavior matches predictive coding predictions but would require different physiological implementations at each contrast level, suggesting the filter shape is functionally optimized rather than a fixed biophysical artifact.Chklovskii introduces the lattice filter as a specific circuit implementation where decorrelation occurs in hierarchical stages, each operating at a different timescale. This architecture predicts that LGN temporal receptive fields should be longer than retinal ones, which matches electrophysiological observations. It also predicts two distinct LGN cell types corresponding to forward and backward prediction error pathways, identifiable with the known lagged and non-lagged cell classes. At Janelia Farm, his group has reconstructed the connectome of the fly visual system through the first two neuropils, mapping approximately 10,000 synaptic connections among 50 neurons per processing column. The L1 and L2 large monopolar cells show response properties consistent with the dual pathways of a lattice filter, and inter-column connections provide the substrate for motion detection.

What if vision isn't a movie playing in your head but a rapid-fire sequence of information-gathering missions, each lasting a third of a second? Dana Ballard dismantles the saliency map paradigm and reveals how dopamine, uncertainty, and internal agendas govern where your eyes go next.Subscribe for more from the Convergent Science Network podcast series.Ballard opens with a fact most people find shocking: high-resolution binocular vision covers only about one degree of visual angle, roughly the width of a thumb at arm's length. Every third of a second, the eyes jump to a new fixation point, meaning vision is fundamentally discrete rather than continuous. The dominant saliency map theory proposes that eyes are drawn to visually complex regions, but Ballard champions the agenda-driven alternative: each fixation serves a specific task, extracting a quantum of information that the brain integrates into the experience of seeing. A possible compromise allows agenda-driven saliency, where task demands modulate what counts as interesting in the image.The interview describes virtual reality experiments where subjects walk down a sidewalk performing three simultaneous tasks: picking up litter, avoiding obstacles, and staying on the path. Eye movement analysis reveals which task the brain is working on at each moment, supporting the idea that complex behavior decomposes into small programs executed in rapid succession. Critically, gaze patterns differ depending on the affordance of an object: eyes fixate on edges when navigating around obstacles but on centers when reaching to pick something up, demonstrating that vision serves action rather than building a passive picture.Ballard connects this framework to reinforcement learning and dopamine signaling. He proposes that the brain's internal programs are scored by a common neural currency, analogous to the euro, implemented by dopamine. His former student Nathan Sprague showed that pure reward-seeking produces unstable gaze behavior, but the product of reward and uncertainty reduction is stable and outperforms alternatives. The driving force behind eye movements is primarily uncertainty reduction: John Senders' classic experiment, where a clamshell periodically blocked a driver's vision, viscerally demonstrates that it is the uncertainty about your position, not the reward of seeing, that compels you to look.

What will it take to build a prosthetic hand that your brain accepts as part of your own body? Maria Chiara Carozza describes the frontier of neurorobotics, where artificial limbs must not only move on command but generate the sensory feedback that creates body ownership.Subscribe for more from the Convergent Science Network podcast series.Carozza defines human-robot symbiosis as a relationship where robot and user share the same objectives and are interdependent in performing tasks. Her Neuro-Robotics Research Group at Scuola Superiore Sant'Anna in Pisa develops wearable robots that read non-invasive signals, including surface electromyography, limb movements, eye tracking, and physiological indicators, to infer user intentions without surgical implants. The challenge extends beyond reading intent: the robot must also provide sensory feedback through wearable interfaces that stimulate the skin, enabling the user to perceive the environment through the artificial device.The interview explores the rubber hand illusion as a bridge between neurophysiology and robotics. When visual and tactile signals are correlated, subjects develop body ownership for a rubber hand, experiencing a stab to the fake hand as if it were real. Carozza's team is translating this principle into prosthetic design by embedding vibrotactile stimulators inside the socket interface between stump and artificial hand. By mapping finger contact forces to specific stimulation patterns on the residual limb, they aim to create a learned association that could eventually migrate perceptually to the fingertips, leveraging the brain's remarkable capacity for adaptive remapping.Carozza also describes exoskeletons for post-stroke rehabilitation, where an external articulated structure acts in parallel with the weakened natural limb. This parallel configuration creates a fundamental control problem: two manipulators must share the same goal and move in harmony, or the system fails and the patient rejects it. Success rates below 90% task completion are unacceptable because the remaining failures create frustration and social embarrassment. The iterative design process, driven by direct feedback from amputees and stroke patients, reveals that cosmetic appearance, lightweight construction, eight-hour battery life, and comfortable skin interfaces are as critical as motor performance.

Why does a duck-billed platypus have electroreception, and what does that tell us about how 250 million years of evolution sculpted the six-layered cortex that makes you human? Jon Kaas traces the entire arc of mammalian brain evolution from stem reptiles to primates.Subscribe for more from the Convergent Science Network podcast series.Kaas argues that understanding brain evolution is essential for understanding what we are. All mammals share a six-layered neocortex that evolved from a simpler one-layered dorsal cortex in stem reptiles, a structure likely involved in habituation and short-term memory rather than sensory processing. The transition to six layers gave early mammals extraordinary flexibility: the ability to replicate cortical areas, specialize them for different functions, and modify sensory representations by enlarging what matters most, whether whisker maps, nose representations, or echolocation frequencies.The interview reconstructs the ecological pressures that drove early mammalian brain evolution. Small, nocturnal, and hunted by dinosaurs, the first mammals developed high-frequency hearing through the dissociation of jaw bones into inner ear ossicles, enabling mother-offspring communication at frequencies predators could not detect. Olfaction dominated the forebrain, essential for nocturnal foraging. Sensory hairs, likely the precursors of whiskers, provided tactile information before physical contact, a critical advantage in poor light. Kaas emphasizes that the brain's hyperplasticity, its ability to automatically reorganize when peripheral inputs change, was the key innovation enabling rapid diversification.The primate chapter of this story centers on the shift to diurnal, arboreal life after the dinosaur extinction 60 million years ago. Visual processing expanded massively, with the temporal and occipital lobes growing to cover the midbrain. Eye-hand coordination became critical for catching insects on moving branches, driving the development of grasping forepaws and eventually freeing the hand from the mouth. Social group living, essential for ground-dwelling primates facing predators, drove frontal lobe expansion. Throughout, Kaas stresses that cortical and subcortical structures co-evolved, with changes in cortex driving modifications in spinal cord circuitry, thalamic inputs, and midbrain organization.

How does an ant with a brain smaller than a pinhead navigate miles of desert using visual memories that would be unrecognizable to a human eye? Andy Phillipides reveals the elegant simplicity of insect navigation and why it could outperform GPS-dependent robots in denied environments.Subscribe for more from the Convergent Science Network podcast series.Phillipides explains why studying ants in their natural environment is essential: laboratory stimuli produce fundamentally different neural responses than the real world. Desert ants like Melophorus bagoti are ideal subjects because they are social foragers that learn routes in a single trial, their behavior gives a direct readout of their nervous system, and researchers can track their entire foraging range. Crucially, ants do not use cognitive maps. Their route memories are insulated by context, so an ant placed mid-path while fed will head home, while an empty ant placed at the same spot will head outward.The interview dissects the complementary navigation strategies ants employ. Path integration, combining step counting with a polarized-light compass, provides a baseline homing vector but accumulates errors over distance. Ants compensate by deliberately aiming to one side of the nest, much like sailors using dead reckoning would aim to one side of a port. From the very first trip, ants layer visual memories on top of path integration, using the skyline silhouette of trees against the sky as a robust, stable landmark. Phillipides describes his group's visual compass model, where ants store panoramic snapshots oriented toward their goal and recover heading by rotating on the spot to minimize image difference, a strategy supported by observed saccadic scanning behavior in unfamiliar environments.The computational implications are striking. With poor-resolution compound eyes, minimal memory, and limited processing power, ants achieve remarkably robust navigation. Phillipides argues these bioinspired algorithms could serve UAVs, space exploration, and any platform where GPS is unavailable and computational resources are constrained. Taking panoramic images from ant-level positions has already revealed how radically different the visual world appears at ground level, where most of the visual field is sky and ground, and prominent landmarks simply disappear into the background.

Can a robotic fish patrol harbors for pollution while swimming so quietly it never disturbs the marine life it protects? Huosheng Hu describes building fish robots that evolved from aquarium exhibits to autonomous ocean sentinels, alongside brain-controlled wheelchairs for people who cannot move.Subscribe for more from the Convergent Science Network podcast series.Hu traces his journey from industrial automation to biomimetic underwater robots, sparked when an aquarium needed robotic replicas of fish species that could not legally be displayed. His 60-centimeter robotic fish uses four to five discrete motor segments to replicate the S-wave swimming motion captured from real fish via camera analysis. The design includes a buoyancy system mimicking a fish bladder, a center-of-gravity shifting mechanism for depth changes, and sensors ranging from gyroscopes and accelerometers to obstacle-detecting infrared and flow-measuring antennae. An EU-funded project now deploys these robots to monitor ship oil leaks and pollution in ports up to 30 meters deep, using an underwater ultrasonic positioning system analogous to GPS.The advantages over conventional submarine-style robots are significant: fish-like propulsion disturbs neither the environment nor pollution plumes, offers greater maneuverability in narrow passages, and theoretically exceeds the 60% efficiency ceiling of propeller-driven vessels. Safety features ensure that if the underwater positioning system fails, the fish surfaces to acquire satellite GPS and navigate home autonomously.Hu's parallel research on assistive robotics tackles mobility for people with severe disabilities. His brain-computer interface records EEG signals from the motor cortex as users imagine hand or leg movements, training neural networks to translate these patterns into wheelchair commands. Current systems achieve roughly 70% accuracy with healthy subjects after several hours of training, with online learning algorithms adapting to fluctuations in mental state. The wheelchair's own laser scanners and ultrasound sensors provide a safety layer that overrides human commands when obstacles are detected, ensuring safe operation even if the user falls asleep or sends erroneous signals.

When will robots walk through our doors, cook our meals, and potentially steal our identities? Yoseph Bar-Cohen surveys the frontier of electroactive polymers, biomimetic actuators, and the ethical minefield of human-like machines that could one day be indistinguishable from us.Subscribe for more from the Convergent Science Network podcast series.Bar-Cohen frames biomimetics as a field that touches everything from clothing inspired by spider silk to drills modeled on gophers. His current work spans bioinspired drilling systems and legged rovers designed to climb steep terrain on other planets, drawing inspiration from mountain goats. He highlights the octopus as a particularly fascinating model organism, noting its ability to squeeze through narrow spaces, camouflage itself, and manipulate objects with soft tentacles. Replicating these capabilities requires materials science breakthroughs that remain years away.The interview provides a detailed overview of electroactive polymer technology, the closest engineering analog to biological muscle. Bar-Cohen distinguishes two fundamental categories: field-activated (physics-based) polymers that require high voltage but can sustain position without power, and ionic (chemistry-based) polymers that operate at low voltage but suffer from drying and position drift. Current EAP actuators produce forces in the range of 10 kilograms but remain roughly an order of magnitude weaker and slower than biological muscle. His arm-wrestling challenge between human and EAP-driven arms provides an intuitive benchmark for the technology gap.Bar-Cohen confronts the ethical implications of increasingly human-like robots with striking directness. He envisions scenarios where robotic clones could rob banks with your fingerprints and DNA, where household robots could be hijacked to surveil celebrities, and where military biomimetic birds could cause enemies to kill all real birds. Rather than advocating for bans, he argues that society must develop solutions incrementally, just as it has addressed spam, computer viruses, and automobile safety. The conversation reveals a researcher who sees both the transformative promise and genuine dangers of the coming robot revolution.