iiinova — where your digital self keeps evolving

Reposted by Fabian Tatai

Constantin Rothkopf @c-rothkopf.bsky.social · 03/09/2026

Pouring a cup of coffee looks trivial. It isn’t. Humans effortlessly control nonlinear liquid dynamics across different vessels and speeds—while robots still struggle. In our new preprint, we ask how the human motor system solves this everyday control problem. 1/5

35014

Fabian Tatai @fatatai.bsky.social · 20/04/2026

I am presenting our work this week at #NCMKobe26 @ncmsociety.bsky.social Our poster is during the Thursday/Friday session: **P2-G-182** - _A computational account of tactile suppression during reaching movements_ Dimitris Voudouris @dominikstrb.bsky.social Katja Fiehler @c-rothkopf.bsky.social

0115

Fabian Tatai @fatatai.bsky.social · 03/03/2026

Do you think this connects to your findings from self touch?

110

Fabian Tatai @fatatai.bsky.social · 03/03/2026

Our computational model explains their findings by assuming that the lack of vision increases the uncertainty about sensory feedback. Therefore, state estimation in the movement task should rely more on the predictions of the forward model, i.e. suppression should be stronger.

100

Fabian Tatai @fatatai.bsky.social · 03/03/2026

The same receptors targeted by vibration are relevant for state estimation, see e.g. the references cited on page 4 of our preprint. Our computational model prescribes how much this sensory feedback should be weighted against the predictions of a forward model in the ongoing movement task.

010

Fabian Tatai @fatatai.bsky.social · 28/02/2026

How suppression related to movement-dependent sensory predictions interacts with suppression related to predicted (self-)touch remains, in our view, an open question. Checkout this paper from Katja and Dimitris for some ideas: www.nature.com/articles/s41...

nature.com

Dynamic temporal modulation of somatosensory processing during reaching - Scientific Reports

Scientific Reports - Dynamic temporal modulation of somatosensory processing during reaching

120

Fabian Tatai @fatatai.bsky.social · 28/02/2026

In contrast, the studies you cite measured sensitivity on a static finger that was expected to be touched by the other hand.

110

Fabian Tatai @fatatai.bsky.social · 28/02/2026

However, in our experiment, tactile sensitivity was measured on the moving forearm, where our experimental manipulations should not alter predictions of tactile input because of expected contact with the screen.

110

Fabian Tatai @fatatai.bsky.social · 28/02/2026

(2) We think this differs from the type of tactile suppression described in the studies you mention. Those concern prediction of actual touch, more precisely self-touch. We think that prediction of touch is a major driver of tactile suppression, as shown in self-touch and haptic exploration.

120

Fabian Tatai @fatatai.bsky.social · 28/02/2026

We therefore interpret tactile suppression as a proxy for how much somatosensory feedback is weighted during online sensorimotor control.

220

Fabian Tatai @fatatai.bsky.social · 28/02/2026

Computationally, this has the consequence that these two signal sources are weighted differently depending on their relative uncertainty over time, which leads to different amounts of "suppression" during movement.

120

Fabian Tatai @fatatai.bsky.social · 28/02/2026

The external stimulus itself is not predicted per se. The nervous system combines movement-related sensory input with internal predictions to best estimate its state over time.

110

Fabian Tatai @fatatai.bsky.social · 28/02/2026

(1) We are examining externally generated touch during movement. A substantial body of neurophysiological and behavioral work shows that cutaneous afferents carry kinesthetic information relevant for body state estimation (details in the paper) e.g., physoc.onlinelibrary.wiley.com/doi/full/10....

physoc.onlinelibrary.wiley.com

Cutaneous afferents provide a neuronal population vector that encodes the orientation of human ankle movements

The aim of this study was to analyse the directional coding of two-dimensional limb movements by cutaneous afferents from skin areas covering a multidirectional joint, the ankle. The activity of 89 c...

110

Fabian Tatai @fatatai.bsky.social · 28/02/2026

Hi Konstantina, thank you for being so interested in our work! 🙂

140

Fabian Tatai @fatatai.bsky.social · 27/02/2026

For those wondering: What does cutaneous touch (like our vibrotactile stimuli) have to do with body state estimation? Neurophysiological and behavioral work has shown that cutaneous afferents carry kinesthetic information and contribute functionally to movement detection and execution.

100

Fabian Tatai @fatatai.bsky.social · 27/02/2026

Huge thanks to my collaborators Dimitris Voudouris, @dominikstrb.bsky.social, Katja Fiehler, @c-rothkopf.bsky.social and the excellence cluster "The Adaptive Mind" for providing such a stimulating research environment. 🙏

110

Fabian Tatai @fatatai.bsky.social · 27/02/2026

Together, these results provide a quantitative, normative account of tactile suppression during movement and point to a general principle for how tactile input may be regulated during action.

100

Fabian Tatai @fatatai.bsky.social · 27/02/2026

As internal uncertainty about hand position increases, suppression decreases—consistent with greater reliance on somatosensory input. This makes a purely fixed gating account unlikely, as suppression changes systematically when we explicitly manipulate uncertainty.

120

Fabian Tatai @fatatai.bsky.social · 27/02/2026

Why is touch perceived as weaker during movement? In our new preprint 📝, we examine tactile suppression during reaching. Using optimal control theory, we show that tactile suppression reflects dynamic, uncertainty-dependent integration of forward model predictions and sensory feedback.

2399

Fabian Tatai @fatatai.bsky.social · 02/08/2025

Happy to announce that I am presenting a poster today at #CogSci25: Physical reasoning during motor learning aids people at transferring mass, but not motor control mappings. This is joint work with Dominik Ürüm, @mariaeckstein.bsky.social and @c-rothkopf.bsky.social Find out more at P3-T-192!

083

Reposted by Fabian Tatai

Constantin Rothkopf @c-rothkopf.bsky.social · 01/02/2025

Our latest work on understanding the behavior of bounded agents in more naturalistic tasks accepted at #ICLR2025: Inverse decision-making using neural amortized Bayesian actors, with @dominikstrb.bsky.social @tobnie.bsky.social and @jan-peters.bsky.social based on @tobnie.bsky.social MSc thesis

1258

Fabian Tatai @fatatai.bsky.social · 16/01/2025

Finally, we conducted a systematic model comparison to show that our data can not be accounted for with simpler explanations, implying that participants really accounted for how the friction impacts the puck's sliding behavior. Find this and more in our pre-print: www.biorxiv.org/content/10.1...

biorxiv.org

Intuitive sensorimotor decisions under risk take Newtonian physics into account

The success of our interactions with objects in natural tasks depends on generating motor actions, which are subject to sensorimotor variability, the objects’ physical properties, which are governed b...

020

Fabian Tatai @fatatai.bsky.social · 16/01/2025

We fitted the data with a generative model of sliding based on statistical decision theory, and participants even seem to account for their personal variability. Subjects with higher variability undershoot more:

140

Fabian Tatai @fatatai.bsky.social · 16/01/2025

Statistical decision theory predicts that if people maximize their task outcome but also account for how the friction impacts the puck, they should undershoot the target more the higher the penalty but also the larger the distance. This is also what we find in the data:

140

Fabian Tatai @fatatai.bsky.social · 16/01/2025

Participants controlled the puck's initial velocity, to hit the green area to receive a positive score but avoid a penalty from hitting the red area. This process is subject to signal dependent noise, which is further transformed by the physics of friction, determining the final puck position.

120

Fabian Tatai @fatatai.bsky.social · 16/01/2025

To study whether people can do this, we built this fun setup, where people get to slide a real standard hockey puck at targets, while viewing the scene through virtual reality. This offers a natural way of interacting with objects for the participants, but still allows for experimental control.

120

Fabian Tatai @fatatai.bsky.social · 16/01/2025

1. Sensorimotor-control, subject to inherent uncertainty and variability; 2. Intuitive physical reasoning as object manipulations are subject to physical dynamics; 3. Economic decision-making when actions have outcomes with monetary consequences.

120

Fabian Tatai @fatatai.bsky.social · 16/01/2025

Humans are exceptionally proficient at manipulating objects in natural tasks. This skill requires integrating three cognitive faculties:

120

Fabian Tatai @fatatai.bsky.social · 16/01/2025

Hello Bluesky, with my first post, I would like to announce that my first first-author paper is out now! Have you ever wondered whether people account for Newtonian physics when interacting with objects? Together with @dominikstrb.bsky.social and @c-rothkopf.bsky.social we show that they do!

1122

Fabian Tatai @fatatai.bsky.social · 15/01/2025

Thanks, for your effort, I would really like to be added as well!

010

Fabian Tatai @fatatai.bsky.social · 08/01/2025

Great list! Could you please add me as well?

000