Sign in
April 25, 2024




The goal of the Neurochats seminar aims to bring together young researchers from Barcelona and to encourage interaction among the various research centers in the city and its surroundings. The event counts with the pleasure of free pizza and the excitement of scientific discovery, creating a relaxed and engaging atmosphere for knowledge exchange. The format includes informal talks lasting 45 minutes followed by a 15-minute discussion, allowing masters, PhD students, and postdoctoral fellows to familiarize themselves with their colleagues’ research.

next session

Date: April 25th, 2024
Time: 18h
Place: UPF, ciutadella campus, Mercè Rodoreda building 24, room 24.S01. MAP

Ring Attractor dynamics underlies the learning of stable working memory representations in a dual task


Working memory (WM) refers to the part of short-term memory that temporarily holds and processes information when it is no longer accessible to the senses. It relies on briefly storing stimulus features in the activity of neuronal populations. Despite its crucial role in cognitive functions, WM is especially vulnerable due to its limited capacity and stability, especially when facing distractions from internal and external sources. Recent studies have explored the neural mechanisms that safeguard WM. It is proposed that pre- and post-distraction neural activity decomposes into orthogonal subspaces, thus protecting information. However, whether orthogonalization is innate or acquired through learning is unknown, and the network mechanisms supporting it are unclear. Here, we probe WM orthogonalization using calcium imaging data from the mouse prelimbic (PrL) and anterior cingulate (ACC) cortices as they learn to perform an olfactory dual task. The dual task combines an outer Delayed Paired-Association task (DPA) with an inner Go-NoGo task. We examined how neuronal activity reflected the process of shielding the DPA sample information against Go/NoGo distractors. As mice learned the task, we measured the overlap between the neural activity onto the low-dimensional subspaces that encode sample or distractor odors. Early in the training, pre-distraction activity overlapped with both sample and distractor subspaces. Later in the training, pre-distraction activity was strictly confined to the sample subspace, resulting in a more robust sample code. We present a mechanistic insight into how these low-dimensional WM representations evolve with learning in a recurrent neural network model of excitatory and inhibitory neurons with low-rank connections. The model links learning to (1) the orthogonalization of sample and distractor WM subspaces and (2) the orthogonalization of each subspace with irrelevant inputs. We validated (1) by measuring the angular distance between the sample and distractor subspaces through learning in the data. Prediction (2) was validated in PrL through the photoinhibition of ACC to PrL inputs, which turned back late training into early training dynamics. Moreover, our model constrains the network dynamics to a double-well attractor on a one-dimensional ring and suggests that learning optimally adjusts the location of the attractors on this ring. We validated this theoretical prediction by estimating an energy landscape for the recorded neural dynamics. In sum, our study underscores the crucial role attractor dynamics plays in shielding WM representations from distracting tasks.


Alexandre Mahrach

Alexandre Mahrach


I studied Physics at Sorbonne Université in Paris both at the undergraduate and graduate levels. My Ph.D. research was in the laboratory of Dr. David Hansel at Université de Paris. I studied how recurrent connectivity can be assessed from optogenetic manipulations in the mouse neocortex. I am currently working under Prof. Albert Compte’s supervision on understanding the dynamics of working memory representations in mice.








For inquiries about this event please contact the Scientific Events Coordinator Ms. Núria Hernández at​​