Prior Experience Shapes Selectivity within Human Ventral Temporal Cortex - Project Summary Rapid visual recognition is crucial for guiding human behavior in dynamic sensory environments, with specialized regions within the ventral temporal cortex (VTC) being thought to support this function. Neuroimaging studies using fMRI consistently show distinct category-selective areas within the VTC, such as those selective for faces, words, and other objects. However, population-level signals such as those from fMRI and LFP mask the rich diversity of tuning observed at the single-neuron level. Rare recordings of single neurons in human VTC suggest an unexpected degree of specificity, sometimes diverging from broader signals and shaped by individual experience. The mechanisms driving this specificity, and how they relate across scales of measurement, remain unknown. The goal of this project is to identify the factors that determine visual selectivity in human VTC by integrating recordings at the single-unit, LFP, and fMRI level in the same subjects. We will specifically test whether visual, semantic, and memorability features explain the observed selectivity, and how prior experience influences neuronal responses. Leveraging a novel methodology to record and track individual neurons, we will dynamically explore large stimulus domains contingent on single neuron responses, allowing for efficient mapping of selectivity profiles. In Aim 1, we use controlled unfamiliar grayscale stimuli (fLoc) that span 10 visual categories and systematically sample relevant stimulus spaces. In Aim 2, we expand the stimulus space to include naturalistic and autobiographically meaningful images (e.g., family, familiar people, places, and events) to examine how real-world experience modulates neuronal responses. Data will be analyzed using shared visual and semantic models, and compared across fMRI, LFP, and single unit to determine how VTC selectivity manifests across scales. Preliminary data demonstrate striking examples of neurons responsive to specific people, concepts, or personal memories, often without a corresponding population-level response. We hypothesize that single- neuron selectivity reflects idiosyncratic combinations of perceptual and experiential features, only partially captured by current models. This work is innovative in its integration of multi-scale measurements in the human brain and its use of dynamic, model-driven stimulus selection. Expected outcomes include new insights into how experience shapes VTC responses, how different measurement scales reflect neural coding, and how functional neuroanatomy relates to the content of neural representations. By integrating multiscale neurophysiology, computational models, and personally relevant stimuli, this project will advance our understanding of visual processing, memory, and the nature of category selectivity in the human brain. These findings may also shed light on altered visual recognition processes in psychiatric conditions such as autism and schizophrenia, where social perception and recognition are often impaired.
