Back

Decision signals in the absence of spiking activity in primate visual cortex

Hou, Y. S.; Liu, L. D.; Pack, C. C.

2025-08-01 neuroscience
10.1101/2025.07.29.667496 bioRxiv
Show abstract

Fluctuations in single-neuron activity in sensory cortex often correlate with perceptual decisions. This kind of correlation has been hypothesized to reflect a causal influence of sensory signals on decisions, a feedback influence of decisions on sensory signals, and various other factors as well. To disentangle these different possibilities, we have examined local field potentials (LFPs) recorded from the middle temporal (MT) area of non-human primates performing a motion discrimination task. Compared to single-neuron spiking, LFPs have the advantage of being decomposable into different frequencies that are associated with different anatomical sources of input. More importantly, they persist when spiking activity is inactivated, which precludes a causal influence of the corresponding neural activity on behavior. We found that high frequency (70-150 Hz) LFP power was correlated with perceptual decisions and that this correlation disappeared when spikes were inactivated, consistent with a causal role for this frequency band in decision making. These signals overlapped in time with decision signals in lower gamma-band power (30-70 Hz), which persisted after spiking inactivation, suggesting a non-causal, feedback input. Interestingly, lower-frequency LFP signals (5-30 Hz) reflected both impending perceptual decisions and the outcome of previous trials. Our results therefore reveal that neural activity multiplexes different sources of information about perceptual decisions and that these types of information can be estimated reliably from different LFP frequencies.

Published in PLOS Biology (predicted rank #15) · training set

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.