The growth of cognition: Free energy minimization and the embryogenesis of cortical computation.

The assumption that during cortical embryogenesis neurons and synaptic connections are selected to form an ensemble maximising synchronous oscillation explains mesoscopic cortical development, and a mechanism for cortical information processing is implied by consistency with the Free Energy Principle and Dynamic Logic. A heteroclinic network emerges, with stable and unstable fixed points of oscillation corresponding to activity in symmetrically connected, versus asymmetrically connected, sets of neurons. Simulations of growth explain a wide range of anatomical observations for columnar and non-columnar cortex, superficial patch connections, and the organization and dynamic interactions of neurone response properties. An antenatal scaffold is created, upon which postnatal learning can establish continuously ordered neuronal representations, permitting matching of co-synchronous fields in multiple cortical areas to solve optimization problems as in Dynamic Logic. Fast synaptic competition partitions equilibria, minimizing "the curse of dimensionality", while perturbations between imperfectly partitioned synchronous fields, under internal reinforcement, enable the cortex to become adaptively self-directed. As learning progresses variational free energy is minimized and entropy bounded.

Different visual stimuli affect body reorientation strategies during sidestepping.

Sidestepping in response to unplanned stimuli is a high-risk maneuver for anterior cruciate ligament (ACL) injuries. Yet, differences in body reorientation strategies between high- and low-level soccer players prior to sidestepping in response to quasi-game-realistic vs non-game-realistic stimuli, remain unknown. Fifteen high-level (semi-professional) and 15 low-level (amateur) soccer players responded to a quasi-game-realistic one-defender scenario (1DS) and two-defender scenario (2DS), and non-game-realistic arrow-planned condition (AP) and arrow-unplanned condition (AUNP). The AP, 1DS, 2DS to AUNP represented increasing time constraints to sidestep. Selected biomechanics from the penultimate step to foot-off were assessed using a mixed-model (stimuli × skill) ANOVA (P < 0.05). Step length decreased in the defender scenarios compared with the arrow conditions. Support foot placement increased laterally, away from mid-pelvis, with increasing temporal constraints. Greater trunk lateral flexion in the 1DS, 2DS, and AUNP has been associated with ACL injury onsets. Higher level players pushed off closer to their pelvic midline at initial foot contact in the 2DS especially. Higher level perception of game-realistic visual information could have contributed to this safer neuromuscular strategy that, when understood better, could potentially be trained in lower level players to reduce ACL injury risk associated with dangerous sidestepping postures.

An outline of functional self-organization in V1: synchrony, STLR and Hebb rules.

A model of self-organization of synapses in the striate cortex is described, and its functional implications discussed. Principal assumptions are: (a) covariance of cell firing declines with distance in cortex, (b) covariance of stimulus characteristics declines with distance in the visual field, and (c) metabolic rates are approximately uniform in all small axonal segments. Under these constraints, Hebbian learning implies a maximally stable synaptic configuration corresponding to anatomically and physiologically realistic ''local maps'', each of macro-columnar size, and each organized as Möbius projections of a "global map" of retinotopic form. Convergence to the maximally stable configuration is facilitated by the spatio-temporal learning rule. A tiling of V1, constructed of approximately mirror-image reflections of each local map by its neighbors, is formed. The model supplements standard concepts of feed-forward visual processing by introducing a new basis for contextual modulation and neural network identifications of visual signals, as perturbation of the synaptic configuration by rapid stimulus transients. On a long time-scale, synaptic development could overwrite the Möbius configuration, while LTP and LTD could mediate synaptic gain on intermediate time-scales.

Contribution of lateral interactions in V1 to organization of response properties.

We propose a model of self-organization of synaptic connections in V1, emphasizing lateral interactions. Subject to Hebbian learning with decay, evolution of synaptic strengths proceeds to a stable state in which all synapses are either saturated, or have minimum pre/post-synaptic coincidence. The most stable configuration gives rise to anatomically realistic "local maps", each of macro-columnar size, and each organized as Mobius projections of retinotopic space. A tiling of V1, constructed of approximately mirror-image reflections of each local map by its neighbors is formed, accounting for orientation-preference singularities, linear zones, and saddle points-with each map linked by connections between sites of common orientation preference. Ocular dominance columns are partly explained as a special case of the same process. The occurrence of direction preference fractures always in odd numbers around singularities is a specific feature explained by the Mobius configuration of the local map. Effects of stimulus velocity, orientation relative to direction of motion, and extension, upon orientation preference, which are not accounted for by spatial filtering, are explained by interactions between the classic receptive field and global V1.

Simulated electrocortical activity at microscopic, mesoscopic, and global scales.

Simulation of electrocortical activity requires (a) determination of the most crucial features to be modelled, (b) specification of state equations with parameters that can be determined against independent measurements, and (c) explanation of electrical events in the brain at several scales. We report our attempts to address these problems, and show that mutually consistent explanations, and simulation of experimental data can be achieved for cortical gamma activity, synchronous oscillation, and the main features of the EEG power spectrum including the cerebral rhythms and evoked potentials. These simulations include consideration of dendritic and synaptic dynamics, AMPA, NMDA, and GABA receptors, and intracortical and cortical/subcortical interactions. We speculate on the way in which Hebbian learning and intrinsic reinforcement processes might complement the brain dynamics thus explained, to produce elementary cognitive operations.

Toward an integrated continuum model of cerebral dynamics: the cerebral rhythms, synchronous oscillation and cortical stability.

Continuum models of cerebral cortex with parameters derived from physiological data, provide explanations of the cerebral rhythms, synchronous oscillation, and autonomous cortical activity in the gamma frequency range, and suggest possible mechanisms for dynamic self-organization in the brain. Dispersion relations and derivations of power spectral response for the models, show that a low frequency resonant mode and associated travelling wave solutions of the models' equations of state can account for the predominant 1/f spectral content of the electroencephalogram (EEG). Large scale activity in the alpha, beta, and gamma bands, is accounted for by thalamocortical interaction, under regulation by diffuse cortical excitation. System impulse responses can be used to model Event-Related Potentials. Further classes of local resonance may be generated by rapid negative feedbacks at active synapses. Activity in the gamma band around 40 Hz, associated with large amplitude oscillations of pulse density, appears at higher levels of cortical activation, and is unstable unless compensated by synaptic feedbacks. Control of cortical stability by synaptic feedbacks offers a partial account of the regulation of autonomous activity within the cortex. Synchronous oscillation occurs between concurrently excited cortical sites, and can be explained by analysis of wave motion radiating from each of the co-active sites. These models are suitable for the introduction of learning rules-most notably the coherent infomax rule.

Synchronous oscillation in the cerebral cortex and object coherence: simulation of basic electrophysiological findings.

A lumped continuum model for electrocortical activity was used to simulate several established experimental findings of synchronous oscillation which have not all been previously embodied in a single explanatory model. Moving-bar visual stimuli of different extension, stimuli moving in different directions, the impact of non-specific cortical activation upon synchronous oscillation, and the frequency content of EEG associated with synchrony were considered. The magnitude of zero lag synchrony was primarily accounted for by the properties of the eigenmodes of the travelling local field potential superposition waves generated by inputs to the cortex, largely independent of the oscillation properties and associated spectral content. Approximation of the differences in cross-correlation observed with differently moving bar stimuli, and of the impact of cortical activation, required added assumptions on (a) spatial coherence of afferent volleys arising from parts of a single stimulus object and (b) the presence of low-amplitude diffuse field noise, with enhancement of cortical signal/noise ratio with respect to the spatially coherent inputs, at higher levels of cortical activation. Synchrony appears to be a ubiquitous property of cortex-like delay networks. Precision in the modelling of synchronous oscillation findings will require detailed description of input pathways, cortical connectivity, cortical stability, and aspects of cortical/subcortical interactions.