Pooled versus structured inhibition drives competitive stimulus interactions across distinct spatial scales of the superior colliculus
Banerjee, A.; Kothari, N. B.; Mysore, S. P.
Show abstract
Competitive stimulus interactions across the space map in the midbrain superior colliculus (SC) are essential for spatial decision-making. Here, with electrophysiological recordings in intermediate and deep layers of the mouse SC (SCid), we discovered that rules and the spatial profiles of these competitive interactions are different between local, within-RF versus global, across-RF-boundary spatial scales. When one visual stimulus was centered in the receptive field (RF) and a second one was located within the RFs classical inhibitory surround, we found that stimulus interactions followed an averaging rule. This classical surround was spatially restricted, and the strength of inhibition underlying these within-RF interactions decreased with distance from the RF center. By contrast, when the second stimulus was located outside the RF, stimulus interactions followed a divisive rule. Strikingly, this extra-classical inhibitory surround was spatially global, and the strength of competitive inhibition underlying these across-RF-boundary interactions was distance-invariant. Computational modeling revealed that whereas within-RF interactions are well-explained by a divisive normalization-like mechanism driven by pooled inhibition, across-RF-boundary interactions are not, and instead, are well-explained by a winner-take-all-like mechanism driven by structured, donut-like inhibition. Combined, our results offer insights into the mechanistic logic of stimulus competition in the SCid, which critically underlies spatial decision-making in mammals.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- The computation of directional selectivity in the Drosophila OFF motion pathway 98%
- Uncovering circuit mechanisms of current sinksand sources with biophysical simulations ofprimary visual cortex 97%
- A recurrent network architecture explains tectal activity dynamics and experience-dependent behaviour 97%
Similar papers in this journal
Similar papers in this journal
"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.