We are excited to share our work on dynamical constraints on neural population activity, published as a cover article in @NatureNeuro. It was led by Emily Oby, @AlanDegenhart, @ErinnGrigsby, with @aaronbatista and team. (1/n).

@MarkChurchland @mjaz_jazlab @FieteGroup @BrunoAverbeck Thank you to @neuroamyo for a wonderful News & Views piece:. (n/n).

@MarkChurchland @mjaz_jazlab @FieteGroup @BrunoAverbeck Here is the cover image (credit Avesta Rastan):. (8/n)

@MarkChurchland @mjaz_jazlab @FieteGroup @BrunoAverbeck We found that it was difficult for animals to violate their natural activity time courses, thereby providing empirical support for network-level computational mechanisms long-hypothesized by network modeling studies. (7/n)

@MarkChurchland @mjaz_jazlab @FieteGroup @BrunoAverbeck This included asking animals to express the same population activity patterns in the opposite temporal order. (6/n)

@MarkChurchland @mjaz_jazlab @FieteGroup @BrunoAverbeck We used a brain-computer interface (BCI) to encourage animals to override their natural activity time courses. (5/n)

@MarkChurchland @mjaz_jazlab @FieteGroup @BrunoAverbeck We ask to what extent are time courses of the neural population activity (i.e., paths of neural trajectories) “carved out” by constraints imposed by the underlying neural circuity. (4/n)

There has been tantalizing evidence of such principles at play in the brain, including beautiful work from @MarkChurchland, Valerio Mante, @mjaz_jazlab, @FieteGroup, David Anderson, @BrunoAverbeck, and many others. (3/n).

Decades of network modeling, including the pioneering work of John Hopfield and others, have posited that the brain's computational abilities relies on time courses of neural activity. (2/n)

RT @scott_linderman: 📣 The 11th Statistical Analysis of Neuronal Data (SAND11) workshop will be held June 11-13, 2025 in New York at the @F….

We've posted a new paper about how to speed up statistical methods for analyzing multi-area recordings by orders of magnitude:. This work was led by Evren Gokcen, with Anna Jasper (Einstein), Adam Kohn (Einstein), and Christian Machens (Champalimaud).

RT @CMUEngineering: Transformative work happens through collaboration. New research from @CMUEngineering & @PittEngineering validates princ….

RT @NatureNeuro: Using a brain–computer interface to challenge monkeys to override their natural time courses of neural activity reveals th….

RT @curi_ms: I'm excited to share our #NeurIPS2024 paper with.@jsoldadomagrane @SmithLabNeuro @YuLikeNeuro!.We develop a new brain stimulat….

We encourage you to try SNOPS here:. (n/n).

SNOPS can also identify specific combinations of activity statistics that the model cannot reproduce, thereby guiding model development. (5/n)

SNOPS can accurately reproduce mean firing rate, Fano factor, pairwise correlation, and population activity statistics (e.g., % shared variance and dimensionality) of neuronal recordings. (4/n)

SNOPS automatically identifies parameters of a spiking network model to reproduce aspects of neuronal recordings. (3/n)

It is challenging to tune parameters of a large-scale spiking network model to produce desired activity characteristics. (2/n)

We are excited to share our new method, SNOPS (Spiking Network Optimization using Population Statistics), published in.@NatComputSci. It was led by @Shenghao_W, with @cc_huang11, Adam Snyder, @SmithLabNeuro, and @BrentDoiron. (1/n).