Reduced rank regression for neural communication: a tutorial for neuroscientists Reduced rank regression (RRR) is a statistical method for finding a low-dimensional linear mapping between a set of high-dimensional inputs and outputs. In recent years, RRR has found numerous applica...

Bichan Wu (@bichanw.bsky.social) & I wrote a tutorial paper on Reduced Rank Regression (RRR) — the statistical method underlying "communication subspaces" from Semedo et al 2019 — aimed at neuroscientists.

arxiv.org/abs/2512.12467

Efficient mixed representation of active and passive motion in the mouse visual thalamus during natural behaviour During natural behaviour, changes in the visual scene are largely driven by the subject’s own movements, which can be actively generated (e.g., walking) or passively imposed by external forces (e.g., ...

Delighted to share our latest preprint. It's been a long time coming. Thanks to all the authors for their unique contribution and for for their patience. We show how the visual thalamus deals with active and passive head motion in freely moving animals: www.biorxiv.org/content/10.1...

3 Postdoctoral Research Fellows Champalimaud Foundation (Fundação D. Anna de Sommer Champalimaud e Dr.

🚨Job alert🚨

The lab has up to *3 postdoc openings* for comp systems neuroscientists interested in describing and manipulating neural population dynamics mediating behaviour

This is part of a collaborative ARIA grant "4D precision control of cortical dynamics"

euraxess.ec.europa.eu/jobs/383909

Thanks to our funders and to @alleninstitute.bsky.social for the open-access dataset which allowed us to investigate this question.

One possibility is that they are specialised for encoding optic flow (see Saleem 2020 www.sciencedirect.com/science/arti...).

We discuss other features of AM and PM that suggest their suitability for the encoding of locomotion-related optic flow in the discussion.

So why are medial HVAs selectively enhanced for visual speed encoding during locomotion?

Our results therefore reveal how a complex interplay of sensory input and ongoing behaviour differentially shapes the efficacy of sensory processing in mouse HVAs. They also highlight the importance of considering behaviour and context when ascribing functional specialisations to mouse HVAs.

This selective enhancement of medial HVAs during locomotion wasn’t generic, but specific to visual speed encoding - drifting gratings direction encoding changed non-selectively across mouse visual cortex.

We were also able to better decode visual speed from population activity in these same areas during locomotion, underlining that the enhancement of visual speed encoding during locomotion is specific to medial HVAs.

Remarkably, when we compared stationary and locomotion behavioural states we found that locomotion selectively enhanced visual speed tuning in medial HVAs AM and PM, along with V1 and LGN.

We next focused on how visual speed encoding changes between behavioural states. Previous research has shown that visual feature tuning tends to improve during locomotion in mouse V1, but whether such enhancements occur non-selectively across mouse visual cortex remains poorly understood.

We focused on how the mouse visual system encodes visual speed and discovered 4 classes of tuning across cortical and thalamic areas. We investigated the properties of these different tuning classes and how they varied between areas.

Locomotion selectively enhances visual speed encoding in mouse medial higher visual areas Mammalian visual systems are comprised of multiple brain areas with distinct functional roles. Whilst functional specialisations have been proposed in the mouse based on visual feature encoding, the e...

Excited to share a new preprint investigating whether functional specialisations of mouse higher visual areas (HVAs) depend on behaviour.
With @amansaleem.bsky.social

Thread below👇

www.biorxiv.org/content/10.1...

1) Some exciting science in turbulent times:

How do mice distinguish self-generated vs. object-generated looming stimuli? Our new study combines VR and neural recordings from superior colliculus (SC) 🧠🐭 to explore this question.

Check out our preprint doi.org/10.1101/2024... 🧵