Emergence of a contrast-invariant representation of naturalistic texture in macaque visual cortex

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 4, 2025

Sensory stimuli vary across a variety of dimensions, like contrast, orientation, or texture. The brain must rely on population representations to disentangle changes in one dimension from another. To understand how the visual system might extract separable stimulus representations, we recorded multiunit neuronal responses texture images varying along two dimensions: property represented as early retina, and naturalistic statistical structure, that modulates V2 V4, but not V1. We measured sites these 3 cortical areas responded variation both dimensions. Contrast modulated all areas. In presence structure increased contrast sensitivity. Tuning for was strongest V4; tuning dimensions most heterogeneous V4. well populations each area could support linear readout Populations V4 only did find evidence robust representation contrast-invariant. Single neurons cortex respond selectively multiple so signals single cannot distinguish simultaneously neural three hierarchically linked - V1, V2, using varied image structure. used decoding methods information about dimension. areas, be decoded independently Only however, contrast. reason is selectivity much more diverse than V1 V2. This heterogeneity allows faithfully represent independent

Language: Английский

Foundations of visual form selectivity for neurons in macaque V1 and V2

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: March 8, 2024

Abstract Neurons early in the primate visual cortical pathway generate responses by combining signals from other neurons: some downstream areas, within same area, and others areas upstream. Here we develop a model that selectively combines afferents derived population of V1 cells. We use this to account for recorded both V2 neurons awake fixating macaque monkeys stimuli composed sparse collection locally oriented features (“droplets”) designed drive subsets neurons. The first stage computes rectified fixed filters at different scales cover field. second weighted combination these first-stage responses, followed final nonlinearity, with parameters optimized fit data physiological recordings constrained encourage sparsity locality. fitted accounts neurons, capturing an average 43% explainable variance 38% V2. models droplet predict classical stimuli, such as gratings orientations spatial frequencies, well textures spectral content, which are known be especially effective driving less effective, however, selectivity include naturalistic image statistics. pattern — defined their weights over 4 dimensions position, orientation, frequency provides common interpretable characterization origin many neuronal response properties cortex.

Language: Английский

Coordinated Response Modulations Enable Flexible Use of Visual Information

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: July 15, 2024

Abstract We use sensory information in remarkably flexible ways. can generalize by ignoring task-irrelevant features, report different features of a stimulus, and actions to perceptual judgment. These forms behavior are associated with small modulations the responses neurons. While existence these response is indisputable, efforts understand their function have been largely relegated theory, where they posited change coding or enable downstream neurons read out visual cognitive using weights. Here, we tested ideas rich, behavioral paradigm, multi-neuron, multi-area recordings primary cortex (V1) mid-level area V4. discovered that those V4 (but not V1) contain ingredients necessary behavior, but via previously hypothesized mechanisms. Instead, demonstrated precisely coordinated across population such ready access correct flexibly guide without making changes synapses. Our results suggest novel computational role for task-dependent modulations: changing gets area, within it. Significance Natural judgments continuous, generalized, flexible. estimate ripeness piece fruit on continuous scale, judging either mango an avocado even though look very different, judge size same fruit, indicate judgment variety behaviors as speaking writing any many languages. show long processes, surround modulation, motor planning sufficient all aspects natural decision-making. find population, reorient reformat representations so relevant used communication example general principle emerges from activity large populations

Language: Английский