Pouya Bashivan

Pouya Bashivan is a computational neuroscientist who builds neural network models that use memory to solve complex tasks and compares them with neural and behavioral measurements from humans and other animals. He earned engineering degrees in Tehran and a computer engineering PhD at the University of Memphis in 2016, and he coauthored an ICLR 2016 paper on deep recurrent-convolutional networks for EEG. As an MIT postdoctoral researcher with James DiCarlo from 2016 to 2020, he was a lead author of a 2019 Science paper in which a neural network model of the visual system generated images that drove neurons in primate area V4 to higher activity than natural images. After a 2020 postdoctoral fellowship at Mila with Irina Rish and Blake Richards, he became a faculty member in McGill's Department of Physiology, received a William Dawson Scholar award in 2022, and coauthored a 2025 NeurIPS spotlight paper that tests vision-language models on perception, attention and memory. He is an associate professor in McGill's Department of Physiology and an associate academic member of Mila.

Now at: Mila and McGill University

Timeline

2001–2016

Engineering degrees and a PhD

He earns a BSc and an MSc in electrical and control engineering at K.N. Toosi University of Technology in Tehran between 2001 and 2009. He completes a PhD in computer engineering at the University of Memphis in 2016, after starting there in 2012.
2015–2016

EEG learning with deep networks

In 2015, he posts "Learning Representations from EEG with Deep Recurrent-Convolutional Neural Networks" to arXiv with Irina Rish, Mohammed Yeasin and Noel Codella, and it appears at ICLR in 2016. The method turns multi-channel EEG recordings into sequences of topology-preserving multi-spectral images and trains a deep recurrent-convolutional network on them, with significant gains in accuracy on mental load classification.
2016–2020

Postdoc at MIT with DiCarlo

From 2016 to 2020, he is a postdoctoral researcher in the Department of Brain and Cognitive Sciences and at the McGovern Institute for Brain Research at MIT, working with James DiCarlo.
2019

Controlling V4 neurons with images

In 2019, he and fellow MIT postdoc Kohitij Kar are lead authors of "Neural population control via deep image synthesis" in Science, with James DiCarlo as senior author. The team maps neurons in primate visual area V4 to nodes of a neural network model of the visual system and uses the model to generate images that drive chosen neurons to about 40 percent more activity than natural images.
2020

Postdoctoral fellow at Mila

In 2020, he becomes a postdoctoral fellow at Mila, where he works with Irina Rish and Blake Richards.
February 2022

A William Dawson Scholar award

In February 2022, McGill's Department of Physiology congratulates him, as a faculty member, on receiving the William Dawson Scholar award. The award is similar to a Canada Research Chair Tier 2, and he appears in McGill's 2026 cohort for a second term running 2026–2031.
2025

A NeurIPS spotlight on VLMs

In 2025, he coauthors "Caption This, Reason That: VLMs Caught in the Middle" with Zihan Weng, Lucas Gomez and Taylor Whittington Webb, accepted as a spotlight at NeurIPS. The paper tests vision-language models, including GPT-4o, on perception, attention and memory tasks drawn from cognitive science.
As of October 2026

Associate professor at McGill

He is an associate professor in the Department of Physiology at McGill University, principal investigator of the Bashivan Lab, and an associate academic member of Mila. His lab develops neural network models that use memory to solve complex tasks and compares them with neural and behavioral measurements from humans and other animals. In August 2026, the lab lists episodic models of the hippocampus, modelling active vision in humans and working memory representations from fMRI data as topics for new graduate students.