💡We concluded that feedback responses may not only benefit from the additive combination of the expectation signal with the incoming sensory information but also be further refined by the prediction error signal acting as a corrective drive in the synaptic input to the primary sensorimotor cortex.
Sensory expectations and prediction error during feedback control in the human brain https://www.biorxiv.org/content/10.64898/2026.01.19.700321v1
🤯 At the same time, in the BOLD activity recorded in M1 and S1, expectation and sensory input were combined through a subtractive rather than additive mechanism, resulting in a signed prediction error signal.
✅ Again, this result was confirmed when we looked at the LFPs recorded in 🐒.
As soon as perturbations occur, the expectation signal is combined with the incoming sensory input through a simple additive mechanism, which we could detect both in the EMG activity (in humans 💁♀️💁♂️) and in M1 spiking activity (in monkeys 🙉🙈).
🔎 Whether this information improves subsequent sensory processing in S1 or is just epiphenomenal to neural processing in other motor regions is now an open question for future investigations.
💡Confirming human results, the LFPs in S1 showed a clear signature of expectations, indicating that this information reaches S1 as synaptic input without influencing spiking activity.
🔎 To understand this difference, we went back to our previous electrophysiological dataset (🐒) and looked at the local field potentials recorded simultaneously with spiking activity. Like the BOLD signal, the LFPs are more influenced by synaptic input and less sensitive to spiking activity.
👀 Interestingly, however, the BOLD signal also showed information about expectations in S1, which was absent in spiking activity recorded in 🐒.
✔️ Similar to @jonathanamichaels.bsky.social's recent results in monkeys published in Nature (👉https://www.nature.com/articles/s41586-025-09690-9), we found that sensory expectations are represented across cortical motor regions through a simple linear geometry.
🔎 We recorded brain activity using fMRI in human participants to understand how probabilistic information about upcoming mechanical disturbances to the body is incorporated into fast feedback responses.
🚨🧵 Happy to share my new preprint “Sensory expectations and prediction error during feedback control in the human brain” with @gribblelab.org , @andpru.bsky.social , @jonathanamichaels.bsky.social and @diedrichsenjorn.bsky.social.
www.biorxiv.org/content/10.6...
Is that the Gulf of Neuroscience on the left? 🥲
If you are interested in a post-doc studying the neural control of movement in humans or NHPs, check out this program and get in touch soon! www.uwo.ca/research/fun...
In happier news, here a new motor learning paradigm and paper by Ali Ghavampour, @mnlmrc.bsky.social, @andpru.bsky.social, @jonathanamichaels.bsky.social, @jjodx.bsky.social.
www.biorxiv.org/content/10.1...
