Gabriel Loewinger

Machine Learning Research Scientist, NIMH
PhD Biostatistics, Harvard '22.
Opinions are my own, not my employer's.

[NIH Innovates @NIH_Innovates]

NIH researchers developed a powerful method to track how brain-behavior relationships evolve, revealing insights that standard analyses miss. This could reshape how we study neural activity in psychiatric disorders and addiction. https://t.co/rw4lgQNLJ9

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira • We released R and Python packages and user guides. The methods can be applied to other neural data types too!
• Paper: https://t.co/H8KoD4th5Z
• Code and user guides: https://t.co/MBvhdLMDoQ 13/13

[Gabriel Loewinger @GLoewinger]

Want to test the effect of events/behavior at every trial time-point in photometry analyses? Paper with @erjiastats, @LovingerDavid, @fpereira. “A Statistical Framework for Analysis of Trial-Level Temporal Dynamics in Fiber Photometry Experiments.” python and R packages! 1/13

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira FLMM finds effects obscured by standard analyses! For example, FLMM reveals effects that “wash out” when analyzed with AUCs. In published work, Cue Period AUC finds no effects because it averages over time- windows (1) and (2) that have opposing effects. 12/13

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira FLMM can disentangle components with distinct temporal dynamics. It can also be used to run analogues of standard hypothesis tests (e.g., ANOVAs, correlations) at each trial time-point. Below is an example akin to the FLMM version of a paired t-test. 11/13

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira Informally, functional random-effects allow one to model variability across animals in the signal “shape.” 10/13

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira Functional random-effects allow one to model how the dynamics of signal-covariate associations vary across animals. 9/13

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira FLMM plots can be conceptualized as pooling signal values (dF/F) at a given trial time-point (e.g., 1.7 sec) across animals and trials, correlating it with covariate(s) (e.g., Latency-to-press) and plotting the slope of the correlation. 8/13

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira FLMM outputs a coefficient estimate plot that shows how the signal–covariate association evolves across trial time-points. 7/13

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira FLMMs exploit autocorrelation to construct *joint* 95% CIs (light grey) that show time windows where effects are statistically significant (any intervals that do not contain 0). All you need to do is visually inspect! 6/13

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira FLMM combines the benefits of 1) Mixed Models to account for between-animal heterogeneity, and 2) Functional Regression to model effects at each trial time-point. 5/13

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira Solution: We propose an analysis framework based on Functional Linear Mixed Models (FLMM) that allows one to analyze signal– covariate associations at every trial time point. 4/13

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira Problem: Photometry is often applied in nested longitudinal experiments with multiple trials per session and sessions per animal. This induces correlation, missing data, etc., that can obscure effects if not accounted for statistically. 3/13

[Gabriel Loewinger @GLoewinger]

@erjiastats @LovingerDavid @fpereira Problem: Common photometry analysis methods reduce detection of effects because, among other things, they average across trials and use 0 summary statistics (e.g., AUC, peak amplitude). 2/13

[Micah Loewinger @MicahLoewinger]

I am so excited to be co-host of @onthemedia, a show that I've loved making for over 8 years. @otmbrooke and our EP Katya Rogers have been incredible mentors. Thank you to OTM's world-class producers @eloiseblondiau @mollyfication @Rebecca_CC_ @cwango_ Thanks for listening ❤️

[Optogenetics Gordon Research Conference @OptoGRC]

Looking forward to a STELLAR #OptoGRC2024 session on interpreting neural manipulations 🧠 🌟

Featuring speakers Daniel O'Shea, @jamesgheys1, @bingbrunton, @AmitVinograd & @GLoewinger and led by Dmitry Rinberg

We can't wait to hear from this incredible line-up!

[Gabriel Loewinger @GLoewinger]

@sophiebeas Congratulations Sofia! And now for our shameless plug: Sofia's paper provides a great demonstration of how functional mixed models reveals effects obscured by summary measure (e.g., AUC) analyses of photometry data. See our paper at:
https://t.co/YLREYNWTOL

[Dr. Sofia Beas @sophiebeas]

Check out our newly published article from our group highlighting how different neurons in the PVT play distinct roles in motivated behaviors:
https://t.co/wFut83SvlJ

[Gabriel Loewinger @GLoewinger]

Come check out my SfN talk tomorrow Monday afternoon on analyzing trial-level photometry data! I will be talking about joint work with @erjiastats, @LovingerDavid, and @fpereira related to our new pre-print https://t.co/CEZaRhRjWY
#SfN2023

[Gabriel Loewinger @GLoewinger]

•We release a package implementing our framework. The methods can be applied to other neural data types too!

•Interested in learning more? I am giving a talk at SfN Monday 1pm in WCC-201

•Code: https://t.co/MBvhdLMDoQ 13/13

[Gabriel Loewinger @GLoewinger]

Want to test the effect of events/behavior at every trial time-point in photometry analyses?? Pre-print w/ @erjiastats, @LovingerDavid, @fpereira: "A Statistical Framework for Analysis of Trial-Level Temporal Dynamics in Fiber Photometry Experiments" https://t.co/7vsdkdkncF. 1/13

[Gabriel Loewinger @GLoewinger]

FLMM finds effects obscured by standard analyses! For example, FLMM reveals effects that "wash out" when analyzed with AUCs. In published work, Cue Period AUC finds no effects because it averages over time-windows (1) and (2) that have opposing effects. 12/13