CCN 2026 | Keynote: Doris Tsao

Item: CCN 2026 | Keynote: Doris Tsao
Rating: 4
Author: PostSingularity Institute

by Doris Tsao

Life & Natural Systems Neuroscience PSANNeurosciences PSAN5Cognitive and behavioural neuroscience

🎙 Doris Tsao 👥 4K 📅 August 12, 2026 ⏱ 65 min 👁 295 📄 original study 🧭 2026-08-15

Available in: English (current) Français

Keywords

face patchesneural codeconsciousnessbinocular rivalryobject representation

Summary

Doris Tsao’s keynote at CCN 2026 presents her research on how the primate brain represents faces and objects, leading to a new theory of consciousness. She begins by describing the macaque face patch system, a network of six face-selective regions in the inferior temporal cortex. Using a shape-appearance model to parameterize faces, she shows that individual face cells encode basis vectors of a high-dimensional face space, with ramp-shaped tuning curves. This axis code generalizes to all of IT cortex, which is organized as maps of object space. She then discusses dynamic changes in the code over time, with a complete reversal at long latencies. The core of the talk focuses on binocular rivalry experiments using Neuropixels probes, revealing that neural populations in face patches rapidly alternate between states representing the conscious and suppressed percepts, at about 20 Hz. This is the first pure neural signal matching conscious perception. She proposes that consciousness requires stable symbols or variables, and that posterior parietal cortex may implement these via ‘persistent object tracking columns’. She concludes with her vision for Astera Neuro, a large-scale effort to understand consciousness through causal manipulation of neural circuits.

190 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk provides high-value information, presenting novel findings on the neural code for faces and its relation to consciousness. The argumentation is solid, based on rigorous experimental methods and clear logical progression. The use of binocular rivalry and Neuropixels probes allows for direct comparison of neural activity with conscious perception. The claim of rapid state alternation is supported by cross-validation and control analyses. The proposal of stable symbols as key to consciousness is speculative but well-motivated by the data.

Scientific Rigor, Source Quality, Title Accuracy

The talk demonstrates high scientific rigor, with detailed methods and results. Sources are primarily the speaker’s own published and unpublished work, with references to collaborators and prior studies (e.g., Logothetis). The title accurately reflects the content, covering both face processing and consciousness. No comments were provided for analysis.

142 words

Title / Content Match

The title accurately reflects the content, covering face processing and consciousness.

Quality & Reliability

8/10

The talk presents original research from a leading neuroscientist, with detailed methods and results, but some claims are unpublished and not yet peer-reviewed.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

Doris Tsao's lab publications — Mentioned as source of research on face patches and object representation.
Logothetis' binocular rivalry studies — Referenced for classic findings on binocular rivalry in visual cortex.

Concurring Sources

Logothetis' binocular rivalry studies — Supports the use of binocular rivalry to study consciousness.

Contribution & Novelties

The talk presents novel findings on the neural code for faces and its dynamics, and proposes a new theory of consciousness based on stable symbols. The discovery of rapid alternation between conscious and suppressed states in IT cortex is a significant contribution.

Pour aller plus loin :

Binocular rivalry — Key paradigm used in the study.
Face perception — Background on face processing.
Neural correlates of consciousness — Relevant to the theory proposed.

73 words

Radar Profile

The radar profile shows high scores in quantity and quality of information, with a strong technical level and good reliability. This reflects a dense, technical talk with original research.

Reliability 8/10