The Brains and Machines Lab

Mission

Our brain continuously processes complex visual information, constructing a comprehensive understanding of the world around us from fleeting 2D retinal impressions. At the Brains and Machines Lab at BGU, we endeavor to understand human visual cognition by employing neural network modeling and model-driven experiments. Our research utilizes a range of techniques, including psychophysics, eye-tracking, and functional magnetic resonance imaging (fMRI), to probe the mechanisms underlying high-order human vision. By leveraging deep learning, we develop computational hypotheses that could explain the observed behavioral and neural data. We then design and conduct innovative model-driven experiments to empirically test these hypotheses, closing the loop between theory and experiment.

Team

Tal Golan

Principal investigator

golan.neuro@bgu.ac.il

Itamar Avitan

Doctoral student

avitanit@post.bgu.ac.il

Ishai Bleiweiss

Master's student

ishaib@post.bgu.ac.il

Ohad Sheelo

Master's student (co-advised with Ari Pakman)

ohadshee@post.bgu.ac.il

Tomer Ashkenazy

Master's student

ashtomer@post.bgu.ac.il

Maya Zach

Master's student

mayazac@post.bgu.ac.il

Maya Nimkovsky

Master's student (co-advised with Hamutal Slovin, Bar-Ilan University)

mayanim@post.bgu.ac.il

Ohad Kelner

Research assistant

ohadkel@post.bgu.ac.il

Affiliated Researchers and Students

These researchers and students are co-advised by Tal Golan but are primarily based in other research groups.

Ossnat Bar-Shira

Postdoctoral researcher in Hamutal Slovin's lab, Bar-Ilan University

ossnat.barshira@gmail.com

Jacob Maimon Master's student, co-advised with Ari Pakman

Itai Kohn Master's student, co-advised with Guy Laban

Alumni

Selected Publications/Preprints

(For a complete publication record, see Tal Golan's Google Scholar profile.)

Tal Golan, Heiko H. Schütt, Nikolaus Kriegeskorte. "Making models disagree to learn how brains compute." Nature Reviews Neuroscience, 2026. DOI: 10.1038/s41583-026-01070-0. Full text.

Neural networks are powerful tools for expressing hypotheses about how the brain computes, but their flexibility can make very different models appear equally predictive of neural or behavioral responses on standard stimulus sets. In this review, co-authored with Nikolaus Kriegeskorte (Columbia University) and Heiko H. Schütt (University of Luxembourg), we present controversial stimuli as a framework for decisive model comparisons. The review covers the mathematical foundations, connecting controversial-stimulus design to Bayesian optimal experimental design, and surveys studies that selected or synthesized stimuli to make competing models disagree.

Itamar Avitan, Tal Golan. "Model-Behavior Alignment under Flexible Evaluation: When the Best-Fitting Model Isn't the Right One." Advances in Neural Information Processing Systems, 2025. NeurIPS proceedings.

Human perceptual judgments, like picking which of three images stands out, can be predicted from neural network representations with high accuracy. But does that mean the network represents the world as we do? We show that this depends on how you link the model to behavior. If the mapping is too flexible, for example by using many free parameters, it can inflate accuracy and make the wrong model look right. High performance does not always mean alignment between models and humans. Our work offers practical steps toward reliable measures of model-human alignment.

Tal Golan, Prashant C. Raju, Nikolaus Kriegeskorte. "Controversial stimuli: Pitting neural networks against each other as models of human cognition." Proceedings of the National Academy of Sciences, vol. 117, no. 47, pp. 29330–29337, 2020. DOI: 10.1073/pnas.1912334117.

This is a seminal work for our research program. It introduced and employed the method of controversial stimuli, which are stimuli designed to elicit distinct predictions in alternative models. Using this method, we discovered that generative neural network classifiers are more human-like than discriminative neural network classifiers.

Tal Golan*, Matthew Siegelman*, Nikolaus Kriegeskorte, Christopher Baldassano. "Testing the limits of natural language models for predicting human language judgements." Nature Machine Intelligence, 1–13, 2023. DOI: 10.1038/s42256-023-00718-1. Full text.

In this work, we translated the controversial stimuli method to the domain of language models, creating pairs of sentences where one sentence was highly likely according to one language model and highly unlikely according to another model, and vice versa. This manipulation allowed us to effectively compare language models in terms of their validity as cognitive models of human relative acceptability judgments.

Kamila M. Jozwik*, Jonathan O'Keeffe*, Katherine R. Storrs*, Wenxuan Guo, Tal Golan, Nikolaus Kriegeskorte. "Face dissimilarity judgments are predicted by representational distance in morphable and image-computable models." Proceedings of the National Academy of Sciences, vol. 119, no. 27, 2022. DOI: 10.1073/pnas.2115047119.

Similarity and dissimilarity relations among stimuli define a geometry in a representational space. In this work, led by several graduates of Niko Kriegeskorte's old Cambridge lab, the representational geometry of human face similarity judgments is compared to that of multiple computational models, including the Basel Face Model and convolutional neural networks trained on various tasks.

Tal Golan*, Wenxuan Guo*, Heiko H. Schütt, Nikolaus Kriegeskorte. "Distinguishing representational geometries with controversial stimuli: Bayesian experimental design and its application to face dissimilarity judgments." SVRHM 2022 Workshop @ NeurIPS, 2022. openreview.net/forum?id=a3YPu2-Mf2h.

In the previous study, we encountered a challenge: multiple representational models predicted the human similarity ratings equally well. In this work, we introduced a method for generating stimuli that elicit distinct representational geometries in different models. Using the synthetic controversial face stimuli in a human psychophysical experiment, we found that training on "inverse rendering" (estimating structure from pixels) results in a neural network that is more human-like than when training the same architecture on alternative tasks.

Tal Golan, JohnMark Taylor, Heiko Herbert Schütt, Benjamin Peters, Rowan Paolo Sommers, Katja Seeliger, Adrien Doerig, Paul Linton, Talia Konkle, Marcel van Gerven, Konrad Kording, Blake Richards, Tim Christian Kietzmann, Grace W. Lindsay, Nikolaus Kriegeskorte. "Deep neural networks are not a single hypothesis but a language for expressing computational hypotheses." Behavioral and Brain Sciences, vol. 46, 2023, e392. DOI: 10.1017/S0140525X23001553. Preprint: 10.31234/osf.io/tr7gx.

This is a concise statement from various cognitive computational neuroscience labs, explaining how challenges posed by cognitive psychology findings should drive progress in neural network-based cognitive modeling rather than a return to traditional box-and-arrow models.

Amirhossein Farzmahdi, Wilbert Zarco, Winrich Freiwald, Nikolaus Kriegeskorte, Tal Golan. "Emergence of brain-like mirror-symmetric viewpoint tuning in convolutional neural networks." eLife, vol. 13, pp. e90256, 2024. DOI: 10.7554/eLife.90256.

The AL face-patch in macaque monkeys is sensitive to faces' head orientations, but it does not differentiate between mirrored orientations (e.g., left and right profiles). In this study, we show that convolutional neural networks (CNNs) trained to discriminate between real-world objects (which are mostly bilaterally symmetric) provide a parsimonious explanation for this property.

Guy Gaziv*, Roman Beliy*, Niv Granot*, Assaf Hoogi, Francesca Strappini, Tal Golan, Michal Irani. "Self-supervised Natural Image Reconstruction and Large-scale Semantic Classification from Brain Activity." NeuroImage, vol. 254, pp. 119121, 2022. DOI: 10.1016/j.neuroimage.2022.119121.

In this cool collaboration with Michal Irani's lab, self-supervised neural networks are trained to reconstruct seen images from fMRI signals. The self-supervision "trick" allows the networks to learn also from images for which no fMRI data is available.

Nikolaus Kriegeskorte, Tal Golan. "Neural network models and deep learning." Current Biology, vol. 29, no. 7, pp. R231–R236, 2019. DOI: 10.1016/j.cub.2019.02.034.

This is an accessible tutorial on neural network models and deep learning.

Openings

• • Postdoctoral opportunities: We welcome inquiries from talented early-career researchers who have completed their PhD or expect to complete it within the next two years. Several programs offer fully funded fellowships for researchers relocating to Israel. We work closely with prospective candidates to identify suitable programs and prepare competitive applications together. Programs include the Azrieli International Postdoctoral Fellowship (upcoming deadline: November 11, 2026), EMBO Postdoctoral Fellowships (upcoming deadline: January 22, 2027), Marie Skłodowska-Curie Postdoctoral Fellowships, Fulbright Postdoctoral Fellowship, Human Frontier Science Program (HFSP) Postdoctoral Fellowships, and the Zuckerman STEM Postdoctoral Scholars Program.
  • Doctoral and Master's opportunities: Applications for the school year starting in October 2026 are closed; we are now considering Master's and PhD students for the school year starting in October 2027.
  • Undergraduate students: Recruitment for the 2026-2027 academic year is closed. Applications for the 2027-2028 academic year will open in March 2027.

  To inquire, please send your CV, current grade sheets, and a brief statement of research interests to golan.neuro@bgu.ac.il.

Contact