Hendrikx, E. | Search bioRxiv

bioRxiv2024

Numerosity adaptation suppresses early visual responses

Humans and many animals rapidly and accurately perceive numerosity, the number of objects, in a visual image. The numerosity of recently viewed images influences our perception of the current images numerosity: numerosity adaptation. How does numerosity adaptation affect responses to numerosity in the brain? Recent studies show both early visual responses that monotonically increase with numerosity, and later numerosity-tuned responses that peak at different (preferred) numerosities in different neural populations. We have recently shown that numerosity adaptation affects the preferred numerosity of numerosity-tuned neural populations. We have also shown that early visual monotonic responses reflect image contrast, which follows numerosity closely. Here we ask how monotonic responses in the early visual cortex are affected by adaptation to different numerosities, using ultra-high field (7T) fMRI and neural model-based analyses. FMRI response amplitudes increased monotonically with numerosity throughout the early visual field maps (V1-V3, hV4, LO1-LO2 & V3A/B). This increase in response amplitudes becomes less steep after adaptation to higher numerosities, with this effect becoming stronger through the early visual hierarchy. This suppression of responses to numerosity is consistent with perceptual effects where adaptation to high numerosities reduces the perceived numerosity. These results imply that numerosity adaptation effects in later numerosity-tuned neural populations may originate in early visual areas that respond to image contrast in the adapting image.

neuroscience↗

bioRxiv2024

Transitions from monotonic to tuned responses in recurrent neural network models during timing prediction

The brain exhibits a gradual transition in responses to visual event duration and frequency through the visual processing hierarchy: from monotonically increasing to timing-tuned responses. Over their hierarchies, properties of both response types are progressively transformed. Here, we implement simulations based on artificial neural networks to investigate the requirements of neural systems for the emergence of such responses and their properties transformations. We see that recurrent networks develop monotonic responses whose properties progressions over network layers resemble those over brain areas. Responses to another sensory quantity, Furthermore, recurrent networks can further develop tuned responses, but only with training, a gradual transition between monotonic and tuned responses emerges. Particularly, if this training is done on predictable sequences, the tuned properties progressions resemble those observed in the brain. These results suggest that the emergence of visual timing-tuned responses and the subsequent hierarchical transformations of these responses result from recurrent neural computation and predictive processing of sensory event timing.

neuroscience↗