Does the brain use all the visual information available in its earliest cortical signals when deciding that something is there?

Details

Perceptual decisions are likely to rely on signals from populations of neurons in early sensory cortical areas. How these responses are combined across neurons and over time to guide behavior is unknown.

To investigate the link between neural responses and perceptual decisions, researchers recorded single-unit (SU) and multi-unit (MU) activity in the primary visual cortex (V1) of monkeys performing a reaction-time visual detection task. They then assessed how well a target could be detected from these neural signals.

On average, SU and MU signals in V1 supported detection sensitivities comparable to the monkey’s detection sensitivity, even when the analysis considered brief response intervals with a median duration of 137 ms that ended shortly before the monkey’s reaction time. However, the overall shapes of the neurometric functions, which describe neural detection performance, differed systematically from the monkey’s psychometric functions, which describe behavioral detection performance.

The researchers also examined the quantitative relationship between SU and MU activity: MU responses were consistent with the summed responses of multiple single units, most of which had low stimulus selectivity. V1 activity and behavioral choices showed weak but significant trial-to-trial covariation, demonstrating for the first time that choice probability could be observed at the earliest stages of cortical sensory processing.

Together, the findings suggest that subsequent processing stages combine the activity of a large population of V1 neurons suboptimally to produce behavioral performance in visual detection tasks.