Really excited to announce that I'll be starting at McGill @mcgillu as an Assistant Professor in the Department of Psychology @McGillPsych in January 2024! .My lab will combine theoretical and systems neuroscience to explore the structure of distributed neural representations

RT @anna_beyeler: 👨‍💻 Open PI position in our institute 👩‍💻 !! If you are an expert in Computational Neuroscience and want to start your la….

RT @biorxiv_neursci: Mapping the projectional architecture of the mouse midbrain dopaminergic system using cell type-specific barcoding htt….

The paper can be accessed freely at

To conclude our study establishes a new paradigm to understand the functional role of prediction error computation in dopamine neurons and opens avenues to develop mechanistic explanations for deficits in intertemporal choice in disease and inspire the design of new RL algorithms.

By recording single neurons across both tasks, we show that the discount factor at the single neuron level is correlated across tasks suggesting that discounting is a cell-specific property, constraining implementations of dopamine-based multi-timescale RL.

Next, we show that heterogeneity in ramping activity across dopamine neurons when mice approach a reward in a 1-D VR track can also be understood as a signature of diverse discount factors across neurons.

First, we quantify the discount factor by measuring the relative response to cues predicting rewards at different delays and show that reward timing information can be decoded from the vectorized prediction error.

Next, we show in two behavioral tasks that single dopamine neurons exhibit a diversity of discount factors.

We first explore 4 computational advantages of multi-timescale RL agents:.- disentangling reward timing and reward magnitude.- learn values with non-exponential temporal discounts.- infer temporal information before convergence.- implement a state-dependent discount factor

Human and animals tend to discount hyperbolically. Far from being irrational, non-exponential discounting can be optimal depending on the uncertainty in the environment and arises in agents combining multiple discounting timescales.

In reinforcement learning (and decision-making in general), the discount factor controls how much a future value is discounted by. Typically agents discount future rewards exponentially according to a single time scale.

Together with this paper by the group of Joe Paton our results across the two papers provide strong evidence that distinct dopamine neurons discount future rewards at different rates providing a substrate for multi-timescale RL in the brain.

Our work with @pablo_tano8 ,.@HyungGoo_Kim @AtharNMalik @pouget_alex .and Nao Uchida exploring how dopamine neurons could enable multi-timescale reinforcement learning in the brain is out in @Nature .

RT @ninamiolane: 🌟 What an opportunity to do exceptional research with an exceptional PI!.

RT @BTolooshams: I am joining @UAlberta as a faculty member and @AmiiThinks this June! Excited to build my research group, working at the i….

RT @BTolooshams: We have released VARS-fUSI: Variable sampling for fast and efficient functional ultrasound imaging (fUSI) using neural ope….

RT @HBHLMcGill: Don’t miss the panel on computational neuroscience at #HBHLSymposium2025 on May 7, 2025! Join experts from @mcgillu, @CAMHn….

Blog post by @dunbar_ba , @SaraM306 and @BTolooshams on our recent @NeuroCellPress paper proposing a new deconvolutional method based on algorithm unrolling to anlayze neural data across recording modalities.

RT @KepecsLab: 1/ Why do late-stage cancer patients lose motivation & sink into apathy?.🔥Our new Science paper shows cancer activates a cyt….

RT @RichardSSutton: The PhD thesis of my _first_ PhD student, Doina Precup, is at-long-last available in digital form. Title: Temporal Abs….