Alan Wrench
@awrench.bsky.social
180 followers 38 following 48 posts
Developing instrumentation for imaging the tongue. Neuroanatomy and biomechanical models. Pulse-Step model of motor control. Natural sceptic.
Posts Media Videos Starter Packs
Reposted by Alan Wrench
avsp.bsky.social
Temporal integration in human auditory cortex is predominantly yoked to absolute time www.nature.com/articles/s41... There's a difference between integrating across absolute time & structure (say) phonemes. Do cortical computations reflect time or structure? Results showed time-yoked computations ⏱️
Temporal integration in human auditory cortex is predominantly yoked to absolute time - Nature Neuroscience
Temporal integration throughout the human auditory cortex is predominantly locked to absolute time and does not vary with the duration of speech structures such as phonemes or words.
www.nature.com
awrench.bsky.social
Note all the articulation happening in the posterior tongue and hyoid which is not captured by EMA.
awrench.bsky.social
Co-registered EMA and ultrasound. From top left: Ultrasound with tongue contour, Ultrasound keypoints, 3D head with EMA sensors, , Spectrogram, Glossogram showing vocal tract constrictions in red cavities in blue, waveform. Movie created by AAA app. Use settings to hear audio.
Reposted by Alan Wrench
thueber.bsky.social
🚨 Open PhD Position – Grenoble, France 🚨

Join us at GIPSA-lab to explore how Speech Language Models can learn like children: through physical and social interaction. Think AI, robots, development 🧠🤖🎙️
Fully funded (3 yrs) • @cnrs.fr / @ugrenoblealpes.bsky.social
Details 👉 tinyurl.com/bde988b3
tinyurl.com
awrench.bsky.social
I understand that you need to minimize kit to take to remote parts. A mirror would be convenient. I think we need to do two things. Change the 50mm camera for a wide angle one so that a peripheral mirror is in view. Then design a 45° mirror mount to fit on the side camera mount. I can try this out.
awrench.bsky.social
Hi Matt, AAA can only record one video channel for reasons to do with having to associate splines with input data streams and also the large amount of disk space that would accumulate. In the past we used a CCTV camera mixer. However, I purchased one recently and couldn't get it to work.
awrench.bsky.social
x.com/i/status/195...

Video of Prof Takayuki Arai with his vocal tract models at #Interspeech2025
Love this analogue demonstration.
Reposted by Alan Wrench
jonathanamichaels.bsky.social
We're thrilled to introduce ATHENA: Automatically Tracking Hands Expertly with No Annotations – our open-source, Python-based toolbox for 3D markerless hand tracking!

Paper: www.biorxiv.org/content/10....
awrench.bsky.social
Really? github.com/HKUDS/DeepCode
www.youtube.com/watch?v=PRgm...
Paper2Code: Convert research papers into working implementations
Text2Web: Generate frontend applications from descriptions
Text2Backend: Create scalable backend systems automatically
Auto Code Validation: Guaranteed working code
DeepCode
YouTube video by Zongwei Li
www.youtube.com
awrench.bsky.social
The work coming out of the Person lab is a must-read for me. www.biorxiv.org/content/10.1...

This new paper shows that cerebellar output neurons encode both predictive and corrective movements, mechanistically linking feedforward and feedback control.
awrench.bsky.social
Although I no longer subscribe to the Equilibrium Point Hypothesis as currently formulated, this proposal is intriguing.

Almanzor et al. (2025). Self-organising bio-inspired reflex circuits for robust motor coordination in artificial musculoskeletal systems.
iopscience.iop.org/article/10.1...
awrench.bsky.social
Not inconsistent with the proposal that an initial feedforward motor cortical pulse determines direction and velocity of movement and a step change near peak velocity modulates deceleration to bring the movement on target. This paper, finds that the cerebellum generates the deceleration step change.
awrench.bsky.social
I believe the cerebellum and possibly motor nuclei learn and map the expected sensory input for a given set of feedforward muscle activations. If there is a mismatch, the cerebellum generates corrective output. If the mismatch persists the cerebellum slowly adapts to the new sensory expectation.
awrench.bsky.social
Nice paper but overlooks the groundbreaking and elegant work describing movement sequence learning. Sosnik, R., Hauptmann, B., Karni, A., & Flash, T. (2004). When practice leads to co-articulation: the evolution of geometrically defined movement primitives. Experimental Brain Research, 156, 422-438.
awrench.bsky.social
Video corresponding to above glossogram. Red line is midsagittal tongue contour automatically estimated using #DeepLabCut Blue line indicates base of mandible to hyoid and purple line indicates base of mandible to short tendon. In collaboration with @drjoanma.bsky.social
awrench.bsky.social
Glossogram with dark red indicating constriction and blue diagonal (tongue compartment contracted) demonstrating peristaltic transfer of water bolus from oral-pharyngeal. This is easiest to explain as sequential extension of neuromuscular compartments of the tongue.
awrench.bsky.social
Great to see this out. Will read it carefully.