TUTORIALS
COSYNE 2026 Tutorial Session Sponsored by the Simons Foundation
TOPIC: Comparative Analysis of Neural Population Codes
SPEAKER: Alex Williams, New York University/Flatiron Institute Center for Computational Neuroscience
DATE: 12 March 2026
TIME: 1:00 p.m. - 5:00 p.m.
Large-scale neural recordings across diverse organisms, alongside an expanding array of deep and recurrent network models, present a key challenge: how to systematically quantify similarities and differences in the population dynamics and representations of these systems. Recent approaches—from linear predictivity and canonical correlation analysis to representational similarity analysis and Procrustes distance—reflect a proliferation of strategies without clear unification.
This tutorial will introduce a framework showing that many of these comparison methods share common mathematical principles, suggesting a path toward standardized tools. As comparative cross-system analyses underpin progress in many other areas of biology—anatomy, physiology, evolution, and genetics, to name a few—a unified approach to comparing neural population codes promises to accelerate discoveries in computational neuroscience.
Please keep in mind that the tutorial has a limited number of participants and that a separate registration fee is required to attend. When registering for the main meeting, select the COSYNE Tutorial add-on.
Alex Williams
New York University/Flatiron Institute Center for Computational Neuroscience
SPEAKER BIOGRAPHY:
Alex Williams is Assistant Professor of Neural Science at New York University and a Project Leader, Associate Research Scientist at the Flatiron Institute Center for Computational Neuroscience. His lab develops statistical methods for analyzing large-scale neural and behavioral measurements, with a special focus on methods for characterizing trial-to-trial and animal-to-animal variability in neural population activity.
COSYNE Tutorials are sponsored by the Simons Foundation. Past COSYNE Tutorials can be found here.
Building AI-ready Data Workflows for Neuroscience Experiments
SPEAKERS: Dimitri Yatsenko, PhD amd Milagros Marín, PhD, DataJoint Inc.
DATE: 12 March 2026
TIME: 8:00 a.m. - 9:00 a.m.
Modern neuroscience experiments—from electrophysiology to multiphoton imaging to behavior—generate complex datasets that demand systematic management and analysis. This hands-on tutorial introduces DataJoint, an open-source platform that automates data workflows so researchers can focus on science rather than infrastructure.
We will demonstrate ready-to-use open-source pipelines for common experimental modalities:
Array electrophysiology: Neuropixels data with automated Kilosort spike sorting
Calcium imaging: Two-photon and widefield microscopy with cell segmentation
Behavior analysis: Pose estimation (DeepLabCut, SLEAP) and action segmentation (MoSeq)
A key focus is multimodal integration—combining multiple instruments in single experiments and synchronizing data and analysis across modalities. Formal pipelines enable unified queries that span electrophysiology, imaging, and behavior simultaneously.
We will draw on prominent research projects including MICrONS, AEON, and Utah Organoids. With rising demand for non-animal studies, we will highlight workflows for organoid research, demonstrating how the same infrastructure scales from traditional animal models to emerging in vitro systems.
Building these pipelines accomplishes critical research objectives: centralized data management (cloud or on-premises), automated quality control across all modalities, computational automation that scales with your experiments, and streamlined data sharing and publishing with collaborators and journals.
The platform handles technical complexity—relational databases, compute orchestration, containerization—automatically. Participants will interact with pipelines through graphical dashboards (Plotly Dash) and an AI agent interface for conversational data queries. Even researchers without Python experience can leverage standardized pipelines, while those comfortable with code can extend and customize them.
This fast-paced tutorial provides comprehensive follow-on materials for self-guided study and lab adoption.
Prerequisites: Laptop with browser access. Python knowledge helpful but not required.
When registering for the main meeting, click SELECT to add this tutorial. No additional fee is required.
Open Neurophysiology Data Tutorial: NWB, DANDI, and IBL
DATE: 12 March 2026
TIME: 10:30 a.m. - 12:30 p.m.
10:30 a.m. –11:30 a.m. - Introduction to NWB and DANDI
SPEAKER: Ben Dichter, CatalystNeuro
Neurodata Without Borders (NWB) is a data standard for neurophysiology that enables interoperable and performant data storage across labs. The DANDI Archive is a free platform for sharing and collaborating on neurophysiology data, now hosting over 450 NWB datasets from efforts including the Allen Institute and the International Brain Laboratory. This session will cover the tools available for converting and sharing your data, as well as how to find and analyze existing datasets to incorporate into your research.
11:30 a.m. –12:30 p.m. - IBL Open Data Platform and Tools
SPEAKER: Gaelle Chapuis, International Brain Laboratory
In this one-hour tutorial, we will introduce participants to the IBL's open-data platform and its associated software ecosystem. Attendees will learn how to access, navigate, and analyze rich datasets spanning behavior, electrophysiology, and imaging, and how to integrate these resources into their own experimental and computational pipelines. The session will combine live demonstrations with hands-on examples in Python, highlighting best practices for reproducibility and collaborative tool development.
The tutorial is intended for both experimentalists seeking ready-to-use tools for data collection and management, and computational neuroscientists looking for large-scale, standardized datasets to test and develop models of brain function.
When registering for the main meeting, click SELECT to add this tutorial. No additional fee is required.