Stefano Fusi
 | The topic of this article may not meet Wikitia's general notability guideline. |
Stefano Fusi | |
|---|---|
| Born | Florence, Italy |
| Nationality | Italian-American |
| Education | Sapienza University of Rome (Physics) Hebrew University of Jerusalem (PhD) |
| Known for | Mixed selectivity, representational geometry, synaptic mechanisms for memory consolidation |
| Scientific career | |
| Fields | Neuroscience, theoretical neuroscience |
Stefano Fusi is an Italian-American neuroscientist. He is a Professor of Neuroscience at Columbia University and an investigator in the Zuckerman Mind Brain Behavior Institute. His research spans the interface of theoretical and experimental neuroscience, with a focus on understanding how biological neural networks support flexible cognition, memory, and learning. Fusi is best known for his work on mixed selectivity, representational geometry, and synaptic mechanisms for memory consolidation.
Early life and education
Stefano Fusi was born in Florence, Italy.[1] He earned a degree in physics from Sapienza University of Rome in 1992.[2] He obtained a PhD in physics from the Hebrew University of Jerusalem in 1999.[3] After his PhD, he held postdoctoral positions at the University of Bern and Brandeis University.[3] In 2005, he was appointed as an Assistant Professor at ETH Zurich (Swiss Federal Institute of Technology).[1]
Career and appointments
In 2009, Fusi joined Columbia University’s Department of Neuroscience as an Associate Professor.[1][2] He is affiliated with the Mortimer B. Zuckerman Mind Brain Behavior Institute and the Center for Theoretical Neuroscience at Columbia,[3] where he directs a research group focused on computational models of cognition, neural representation, rule learning, and memory.[4][3] Fusi also serves as an associate editor for journals including Journal of Computational Neuroscience, and Neural Computation.[2]
Since 2024, Fusi has been a co-director of the Methods in Computational Neuroscience course[5] at Marine Biological Laboratory in Woods Hole, MA.
Research contributions and interests
Fusi helped advance the concept of nonlinear mixed selectivity,[6][7] showing that neurons that encode combinations of task-relevant variables in a nonlinear fashion can dramatically expand the representational capacity of neural networks. This high-dimensional encoding has been proposed as a critical substrate for context-dependent and flexible behaviors. His work in this area has informed both neuroscience and artificial intelligence, offering a computational rationale for the apparent "mixed" tuning observed in prefrontal cortex and other association areas.
Fusi has been a pioneer in the emerging field of representational geometry,[8] examining how population activity organizes into structured manifolds and how abstract cognitive variables are encoded in subspaces that are orthogonal to sensory input dimensions. His work explores how such geometries evolve with learning and how they support generalization and abstraction—key features of primate cognition.
He has also developed influential theoretical models of memory storage,[9][10] emphasizing the role of synaptic complexity and metaplasticity. His"cascade model of synapses with multiple hidden states provides a framework for understanding how biological networks can achieve long memory lifetimes while retaining plasticity. These models have helped reconcile biological constraints with the need for stability in cognitive systems.
In collaboration with physicists and engineers, Fusi has also worked on applications of neuroscience principles to neuromorphic computing.[11][12] His work has inspired hardware systems that mimic the dynamics of biological synapses and neurons to implement brain-like computation in energy-efficient ways. These contributions aim to bridge the gap between neuroscience and next-generation computing architectures.
References
- ↑ 1.0 1.1 1.2 "Stefano Fusi". Simons Foundation. 2017-07-13. Retrieved 2025-10-14.
- ↑ 2.0 2.1 2.2 "Stefano Fusi". The Data Science Institute at Columbia University. Retrieved 2025-10-14.
- ↑ 3.0 3.1 3.2 3.3 "Stefano Fusi". zuckermaninstitute.columbia.edu. 2017-03-06. Retrieved 2025-10-14.
- ↑ "Stefano Fusi, PhD". Vagelos College of Physicians and Surgeons. 2017-06-12. Retrieved 2025-10-14.
- ↑ "Methods in Computational Neuroscience | Marine Biological Laboratory". www.mbl.edu. Retrieved 2025-10-14.
- ↑ Rigotti, Mattia; Barak, Omri; Warden, Melissa R.; Wang, Xiao-Jing; Daw, Nathaniel D.; Miller, Earl K.; Fusi, Stefano (2013). "The importance of mixed selectivity in complex cognitive tasks". Nature. 497 (7451): 585–590. Bibcode:2013Natur.497..585R. doi:10.1038/nature12160. ISSN 1476-4687. PMC 4412347. PMID 23685452.
- ↑ Fusi, Stefano; Miller, Earl K; Rigotti, Mattia (2016-04-01). "Why neurons mix: high dimensionality for higher cognition". Current Opinion in Neurobiology. Neurobiology of cognitive behavior. 37: 66–74. doi:10.1016/j.conb.2016.01.010. ISSN 0959-4388. PMID 26851755.
- ↑ Bernardi, Silvia; Benna, Marcus K.; Rigotti, Mattia; Munuera, Jérôme; Fusi, Stefano; Salzman, C. Daniel (2020-11-12). "The Geometry of Abstraction in the Hippocampus and Prefrontal Cortex". Cell. 183 (4): 954–967.e21. doi:10.1016/j.cell.2020.09.031. ISSN 0092-8674. PMC 8451959. PMID 33058757.
- ↑ Fusi, Stefano; Drew, Patrick J.; Abbott, L. F. (2005-02-17). "Cascade Models of Synaptically Stored Memories". Neuron. 45 (4): 599–611. doi:10.1016/j.neuron.2005.02.001. ISSN 0896-6273. PMID 15721245.
- ↑ Benna, Marcus K.; Fusi, Stefano (December 2016). "Computational principles of synaptic memory consolidation". Nature Neuroscience. 19 (12): 1697–1706. doi:10.1038/nn.4401. ISSN 1546-1726. PMID 27694992.
- ↑ Amit, Daniel J.; Fusi, Stefano (1994). "Learning in neural networks with material synapses". scholar.google.com. pp. 957–982. Retrieved 2025-10-14.
- ↑ "Spike-driven synaptic plasticity: theory, simulation, VLSI implementation". scholar.google.com. Retrieved 2025-10-14.
External links
This article "Stefano Fusi" is from Wikipedia. The list of its authors can be seen in its historical. Articles taken from Draft Namespace on Wikipedia could be accessed on Wikipedia's Draft Namespace.