iiinova — where your digital self keeps evolving

Rebecca Saxe @rebeccasaxe.bsky.social · 20/08/2026

What an incredibly moving and joy-filled 20-year reunion of Saxe-lab past and present. I feel so lucky to have these people be my scientific family.

2800

Rebecca Saxe @rebeccasaxe.bsky.social · 05/08/2026

When we started scanning human infants 13 years ago, it seemed likely that functional responses in infant cortex would be qualitatively different from adults. Instead we keep finding similarities between infants and adults. Even for scene perception, despite the slow development of navigation.

Image of brain activation when infants look at movies of visual scenes, showing activation in three cortical regions: OPA, PPA and MPA.

17313

Rebecca Saxe @rebeccasaxe.bsky.social · 18/05/2026

This job is part of the RC2 funded initiative "Sustainable Transformation for Research Rigor"

100

Rebecca Saxe @rebeccasaxe.bsky.social · 25/04/2025

Infant fMRI was then a technique in its own infancy. The four infants who contributed the majority of the data to that first study were my first child, my second child, my sister’s child, and a post-doc’s child.

140

Rebecca Saxe @rebeccasaxe.bsky.social · 24/04/2025

The question we ask: when do face responses arise in human infant cortex? Neuroscientists and developmental psychologists (my two disciplines) have different intuitions.

Graphic showing two alternative theories of the sequence and timing of development of cortical face responses: slow and sequential, or early and parallel.

15511

Rebecca Saxe @rebeccasaxe.bsky.social · 07/03/2025

So proud to Stand up for science with these amazing scientists @kidsbrains.bsky.social

222016

Rebecca Saxe @rebeccasaxe.bsky.social · 25/06/2024

All three studies replicated the finding that infants look longer at unexpectedly high jumps than expectedly low jumps. Our new workflow is so fast, that we easily collected 134 participants. In-lab and on Zoom we had 20 and 28 participants respectively

Plots showing the average looking time is higher for unexpected than expected trials in the in-lab, zoom, and automated versions of the same experiment.

100

Rebecca Saxe @rebeccasaxe.bsky.social · 25/06/2024

We tested this new workflow by running the same violation-of-expectation study three times using different workflows: 1) in-lab (previously published by Liu & Spelke, 2017), 2) Zoom and 3) "hands-off". We chose the classic action efficiency paradigm from Gergely & Csibra.