Dimitra Maoutsa

@dimma.bsky.social

1425 Followers 757 Following

Theor/Comp Neuroscientist (postdoc) Prev@TU Munich Stoch&nonlin dynamics @TU Berlin&@MPIDS ... Learning dynamics, plasticity&geometry of representations https://dimitra-maoutsa.github.io https://dimitra-maoutsa.github.io/M-Dims-Blog

Schematic illustrating three possible explanations for greater variance concentration after learning. Before learning, two stimulus groups occupy overlapping clouds in a two-dimensional population space, with a dashed magenta line indicating the stimulus-discriminant axis.
After learning, increased variance in the leading mode could reflect task-irrelevant gain amplification, reduced off-axis variability without a change in representation geometry, or alignment of the dominant variance axis with the stimulus-discriminant axis.

Bar plot showing stimulus selectivity, measured as Cohen’s d, for TCA components ordered by incremental variance explained. Rewarded learning after training shows strong selectivity in the leading components, whereas unrewarded experience shows strong selectivity mainly in later, lower-variance components. The dashed horizontal line indicates the threshold for strong selectivity.

Schematic of the experimental setup and our analysis. Mice experienced visual corridors containing leaf or circle patterns, with reward available only for one pattern in the rewarded cohort and no reward in the unrewarded cohort. Large-scale calcium imaging sampled neurons across V1 and higher visual areas before and after learning. Neural activity was organised into neuron × time × trial tensors and analysed with tensor component analysis. Example factors show differences between rewarded and unrewarded mice before versus after learning, including stronger separation of leaf and circle trials after rewarded learning.