Numerosity tuning in human association cortices and local image contrast representations in early visual cortex

Abstract

Many animals use visual numerosity, the number of items in a group, to guide behavior. Neurons in human association cortices show numerosity-tuned responses, decreasing amplitude with distance from a specific numerosity. How are such responses derived from early visual responses? Recent studies show aggregate response amplitudes in human early visual cortex monotonically increase with numerosity, regardless of object size and spacing. This is surprising because numerosity is typically considered a high-level visual or cognitive feature. Here we first use computational modelling of 7T fMRI data to show these monotonic responses originate at the stimulus's retinotopic location in primary visual cortex (V1). Given this location, we then ask whether these monotonic responses can be better described by V1's established response properties. We characterize the Fourier decomposition (into contrast at specific orientations and spatial frequencies) of laboratory numerosity stimuli. This demonstrates that aggregate Fourier power (at all orientations and spatial frequencies) nonlinearly follows numerosity with little effect of item size, spacing or shape: it is proportional to numerosity at a fixed contrast. This nonlinear relationship lets us distinguish predictions of responses to Fourier power and numerosity. Monotonic responses are better predicted by Fourier power, later tuned responses are better predicted by numerosity. Tuned responses emerge after lateral occipital cortex and are independent of retinotopic location. We propose that numerosity's straightforward perception and neural responses reflect its straightforward estimation from early visual spatial frequency domain image representations. Our numerical vision may have built on behaviorally beneficial analysis of spatial frequency in simpler animals.

Visual timing-tuned responses in human association cortices and response dynamics in early visual cortex

“…This suggests that, although the different quantities are not processed using the exact same neural populations, there may be similar computational mechanisms for estimating different quantities in the brain. In both cases, early visual responses monotonically follow numerosity and timing, and tuned responses emerge later 39 . However, the transition from monotonic responses to numerosity to numerosity-tuned responses takes place around the lateral occipital areas, rather than the temporal occipital areas that we see for timing.…”

“…So, timing-tuned neurons may have large receptive fields that the stimulus must fall within 50 . However, analogous numerosity aftereffects in visual perception 52 appear to reflect changes in the responses of numerosity-tuned neurons 26 , 53 and have a similarly limited spatial spread 54 , but numerosity-tuned neurons do not require spatial receptive fields overlapping the stimulus area 32 , 33 , 39 , 55 . Therefore, relationships between the spatial spread of perceptual quantity adaptation effects and the receptive fields of quantity-tuned neural populations may be more complex than they appear: rather than perceptual quantity aftereffects providing clear evidence demonstrating that quantity-tuned responses follow spatial receptive field properties, neural tuning for different quantities 33 , 34 and visual position 32 , 33 , 39 , 55 seem to be independent dimensions of neural responses.…”

“…Computational models for other quantities suggest that excitatory and inhibitory non-linear monotonic responses can be compared to give numerosity-tuned responses 37 , 38 . Indeed, monotonic responses of early visual areas that closely follow numerosity are likely to be transformed into tuned responses to numerosity in lateral occipital areas 39 . A range of motor event timing-dependent neural response profiles, from timing-tuned to monotonic, has been described in macaque premotor cortex 40 , 41 .…”

Derivation of visual timing-tuned neural responses from early visual stimulus representations

“…This suggests that, although the different quantities are not processed using the exact same neural populations, there may be similar computational mechanisms for estimating different quantities in the brain. In both cases, early visual responses monotonically follow numerosity and timing, and tuned responses emerge later 23 .…”

“…Computational models for other quantities suggest that excitatory and inhibitory non-linear monotonic responses can be compared to give numerosity-tuned responses 21,22 . Indeed, monotonic responses of early visual areas that closely follow numerosity are transformed into tuned responses to numerosity in lateral occipital areas 23 . A range of motor event timing-dependent neural response profiles, from timing-tuned to monotonic, has been described in macaque premotor cortex 24,25 .…”