Neural coding of multiple motion speeds in visual cortical area MT

Segmenting objects from each other and their background is critical for vision. Motion speed provides a salient cue for segmentation, yet how the visual system represents and differentiates multiple speeds remains poorly understood. We investigated the neural coding of multiple motion speeds in the primate visual cortex. First, we characterized the perceptual capacity of human and macaque subjects to segment overlapping random-dot stimuli moving at different speeds. We then recorded from neurons in the middle temporal (MT) cortex of macaque monkeys to determine how multiple speeds are represented. We made a novel finding that the responses of MT neurons to two speeds showed a robust bias toward the faster speed component when both speeds were slow (≤ 20°/s). This faster-speed bias emerged early in the neuronal response. It occurred regardless of whether the two speed components moved in the same or different directions, and even when attention was directed away from the receptive field. As stimulus speed increased, the faster-speed bias diminished and eventually disappeared at high speeds. Our finding can be explained by a modified divisive normalization model, in which the weights for the speed components are proportional to the responses of a population of neurons elicited by the individual speeds. We suggest that the neuron population, referred to as the weighting pool, includes neurons with a broad range of speed preferences. We also showed that a classifier can differentiate the responses of MT neurons to two speeds versus a single log-mean speed. We further showed that it was possible to decode two speeds from MT population response, supporting the theoretical framework of coding multiplicity of visual features in neuronal populations. The decoded speeds from MT population can account for the perceptual performance of segmenting two speeds with a large (×4) difference but not a small (×2) difference, raising questions for future investigations. Our findings help to define the neural coding rule of multiple speeds. The faster-speed bias in MT at slow stimulus speeds could benefit important behavioral tasks such as figure-ground segregation, as figural objects tend to move faster than the background in the natural environment.