Penn Grad Talks 2026 | Mengting Fang Mengting Fang, Psychology “Pause, Then Power: How the Brain Decides When to Work Hard” Natural Sciences Presentations | March 27, 2026 Penn Grad Talks (formerly known as Grad Ben Talks) features TE...

📹 7-min talk on this work: vimeo.com/1179294328

(5/5) Our model is general given its information-theoretic formulation, and thus should be directly applicable to other tasks involving evidence encoding and maintenance.

(4/5) We introduce a novel normative Bayesian updating model that is dynamically constrained by optimal performance-effort trade-offs. With two free parameters, the model accurately predicts the rich dependencies of subjects’ accumulation behavior on the evidence schedule.

(3/5) Instead, our results show that evidence accumulation is actively controlled and resource-rational, with dynamics consistent with a computational strategy that continuously balances the informational benefits of incoming evidence against the cognitive costs of acquiring and maintaining it.

(2/5) We address this question with a combination of targeted psychophysical perturbation experiments and computational modeling. Our study uncovers previously unknown properties of evidence accumulation that challenge the standard view of this process as a stationary process of leaky integration.

(1/5) Evidence accumulation is central to decision-making and has been studied extensively across the behavioral and brain sciences. Yet, despite its importance, how the precise temporal structure of evidence shapes the accumulation process has never been systematically investigated.

🎯New preprint!

The resource-rational dynamics of evidence accumulation
www.biorxiv.org/content/10.6...

Evidence accumulation isn’t stationary.
It is dynamically controlled to trade off information gain against cognitive effort.

With Jiang Mao, Tobias Donner @donnerlab.bsky.social, Alan Stocker

Fig. 1. a. Visual and auditory regions of interest (ROIs). b. Responses in a combination of visual (e.g., early dorsal visual stream; Fig. 1a, middle panel) and auditory regions were used to predict responses in the rest of the brain using MVPN. c. In order to identify brain regions that combine responses from auditory and visual regions, we identified voxels where predictions generated using the combined patterns from auditory regions and one set of visual regions jointly (as shown in Fig. 1b) are significantly more accurate than predictions generated using only auditory regions or only that set of visual regions.

I’m excited to share my 1st first-authored paper, “Distinct portions of superior temporal sulcus combine auditory representations with different visual streams” (with @mtfang.bsky.social and @steanze.bsky.social ), now out in The Journal of Neuroscience!
www.jneurosci.org/content/earl...

#VSS2025 Super excited to present my poster “Categorical Representations in Sequential Evidence Accumulation” on Tuesday May 20, 2:45pm-6:45pm at Pavilion 56.440!

Come say hi if you are curious about how we study categorical decision-making and evidence accumulation 🧠👀

Beautiful work!!