News and opinion from Jack Gallant's Cognitive, Computational & Systems Neuroscience Lab at UC Berkeley
Berkeley, CA
Joined September 2011
- Tweets1K
- Following52
- Followers8.9K
- Likes200
Jack Gallant retweeted
#Layer_fMRI_Artifact_of_the_Week In layer-fMRI, the voxel density is variable across depths and curvature. Thus, it’s hard to statistically compare different depths and columns.
The visualization is inspired from @cvnlab
So many hidden gems in his paper biorxiv.org/content/early/20…
Jack Gallant retweeted
A thought provoking (and beautifully done) post on adversarial examples and L2 regularization
thomas-tanay.github.io/post-…
Jack Gallant retweeted
Our paper “A spatial map of onset and sustained responses to speech in the human superior temporal gyrus” can be found online today in @CurrentBiology! We use ECoG and computational methods to describe a new spatial map in STG! Tweet version to follow. cell.com/current-biology/abs…
Jack Gallant retweeted
Replacing p values with out of sample prediction will just be shifting the problem from p-hacking to out-of-sample-hacking. My feeling is that many machine learning papers also do not replicate.
Jack Gallant retweeted
Nice short read on pitfalls of p-value-driven hypothesis testing lucklab.ucdavis.edu/blog/201…
Jack Gallant retweeted
Polyglot Data Science: Python, R, @JuliaLanguage and Fortran cooperating in-process, through @IPythonDev; demo code:
gist.github.com/fperez/5b492…
Language wars among open tools are pointless, they complement each other.
Video/slides of the talk @UCBIDS: bids.berkeley.edu/resources/…
Jack Gallant retweeted
These results look better than the reconstructions from Thomas’s 2009 paper, but I think the increase in model complexity is much larger than the increase in performance. We’re data limited, people! teapawt.net/biorxiv_neursci/status…
Generative Adversarial Networks Conditioned on Brain Activity Reconstruct Seen Images biorxiv.org/cgi/content/shor… #biorxiv_neursci
Jack Gallant retweeted
With a few exceptions at the tails, the formula
h-index = 0.5 * √∑(citations)
holds with astonishing accuracy (96% correlation.) So let's not read too much into the h-index -- If you want it higher, just publish a lot of (better than average) papers.
thebibliomagician.wordpress.…
Jack Gallant retweeted
New nilearn release: easier neuroimaging with machine learning in Python
nilearn.github.io/whats_new.… :
• Connectivity-matrix ordering for plotting
• Fast brain parcellations
• Some bug fixes
Update with "pip install -U nilearn"
Jack Gallant retweeted
Unsupervised discovery of temporal sequences in high-dimensional datasets, with applications to neuroscience biorxiv.org/cgi/content/shor… #biorxiv_neursci
Jack Gallant retweeted
Damning meta-analysis: Simons, D. J., Boot, W. R., Charness, N., Gathercole, S. E., Chabris, C. F., Hambrick, D. Z., Stine-Morrow, E. A. L. (2016). Do ‘brain training’ programs work? Psychological Science in the Public Interest, 17(3), 103-186. teapawt.net/kemmorey1/status/96524…
Still no new data on brain training from Adam Gazzaley. Lots of hype - I’d like to see replication #AAASmtg
Jack Gallant retweeted
This paper is rad --> doi.org/10.1016/j.neuroimage…
"varying the number of subjects was completely inconsequential, whereas the number of data per subject exerted massive effects on model identifiability." h/t @gallantlab
Jack Gallant retweeted
Tried out the new @caseforge custom mold & combined it with @firmmsoftware . So we can get the lowest movement #BOLD #fMRI data ever collected. Also ... you can look cool ... like a Spartan from 300 ... or Magneto
Jack Gallant retweeted
UC Berkeley @BerkeleyDataSci Launches 'Foundations of Data Science' Certificate Program on edX | ow.ly/mnyf30i09NF
Jack Gallant retweeted
Graph theory by and at large ignores everything we know about the brain. Why should a measure based indiscriminately on the neuroiologically irrelevant AAL regions yield any insight? There is functional specialization and dedicated connectivity! biorxiv.org/content/early/20…
Jack Gallant retweeted
Lots of electrodes on a small shaft with everything integrated. buff.ly/2yJ4yVM - this will make semi-chronic physiology ridiculously easier.
Jack Gallant retweeted
Wowza: NVIDIA presents mind-blowing results from incrementally-trained GANs: youtu.be/XOxxPcy5Gr4 the paper: research.nvidia.com/publicat…
Jack Gallant retweeted
RIP John Lisman, one of the greats of neuroscience, a kind & decent man, a bold & original thinker, a Renaissance idea-man, will miss him!
Jack Gallant retweeted
When considering proposals for redefining or abandoning statistical significance, remember that their effects on andrewgelman.com/2017/10/03/…