Meeting Notes
- Date: 2025-10-03
- Time: 10:00 (PT)
- Location: Teams Meeting
- Presentations: @farznaj
Agenda
Farzaneh Najafi, will present a recent paper from her newly formed laboratory.
Title: Interval timing signaling, not temporal predictive processing, shapes cell-type-specific dynamics in PPC and V1
Neural activity following regular sensory events may reflect either elapsed time since the previous event—temporal signaling—or predictions about the timing of the next event—temporal predictive processing (PP). Disentangling these signals is critical for understanding temporal computations in the cortex. To address this, we combined two-photon calcium imaging in layer 2/3 of mouse primary visual cortex (V1) and posterior parietal cortex (PPC) with passive behavioral paradigms in awake mice, involving repeated presentation of audiovisual stimuli at fixed or variable intervals. Neurons clustered into two broad functional classes: stimulus-activated neurons and stimulus-inhibited neurons with ramp-up activity, each showing diverse dynamics. We found that: (1) neural responses adapted immediately to interval changes, (2) responses showed no differences between fixed and jittered ISIs, and (3) the same functional subtypes were present in naïve mice. These results argue against temporal PP in V1/PPC during passive stimulation. Instead, they align with elapsed-time coding via neurons’ heterogeneous rise and decay kinetics. This diversity provides a distributed population code for time, consistent with stimulus-reset attractor dynamics in recurrent networks. Circuit analysis revealed cell-type biases, with SST neurons contributing to fast stimulus-driven responses and VIP/excitatory neurons to slower ramps. Together, our findings demonstrate that V1 and PPC intrinsically encode elapsed time during passive perception without temporal predictive processing, and suggest that temporal predictions may instead emerge in other brain regions or under active behavioral tasks that require movements or decisions.
Meeting Recording
To be added
💬 Start a discussion for this page on GitHub (A GitHub account is required to create or participate in discussions)