
OHBMX
@ohbmequinox
OHBM EquinoX Twitter conference 20/03/2020; 24 hrs of neuroimaging hijinx as Earth makes 1 rotation on its journey around the Sun! ohbmx.org
ID: 1222873370968121345
https://ohbmx.org/ 30-01-2020 13:26:34
675 Tweet
1,1K Followers
754 Following

OHBMX Ev (like in 'evidence', not Eve) Fedorenko 🇺🇦 2️⃣U can use a task to find FUNCTIONAL regions in each individual brain. But to eyeball maps (“blob A in brain X ~ blob B in brain Y”) is unprincipled, & hard if activity is a swath (not blobs). So Ev (like in 'evidence', not Eve) Fedorenko 🇺🇦 devised group-constrained, subject-specific localization (GSS) #OHBMx


OHBMX Ev (like in 'evidence', not Eve) Fedorenko 🇺🇦 3️⃣GSS should be better (shorturl.at/adlNU). So we pitted it against RFX! Tested 5 networks differing in laterality, domain- & modality-specificity/generality: language (n=510 ppl), Multiple Demand (392), Default Mode (386), Theory of Mind (132), face processing (102) #OHBMx


OHBMX Ev (like in 'evidence', not Eve) Fedorenko 🇺🇦 4️⃣First we generated broad masks to constrain fROI location (top Fig). GSS = fROIs may slightly vary in location across ppl as long as within same mask. RFX = group-based fROIs (traditional). For both, effect sizes extracted w/ across run cross validation (CV; bottom Fig.) #OHBMx


OHBMX Ev (like in 'evidence', not Eve) Fedorenko 🇺🇦 5️⃣GSS ALWAYS outperformed RFX, & by a huge margin! Note: even w/ n>100, RFX totally misses some regions (0 activations, see arrows) but GSS finds robust effects; in MD, classic response is "hard>easy", but GSS-easy > RFX-hard. So RFX massively underestimates effects #OHBMx


OHBMX Ev (like in 'evidence', not Eve) Fedorenko 🇺🇦 6️⃣GSS is principled, easy (toolbox: osf.io/wm6zs/) & CHEAP (for our tasks, >80% power w/ n=10). Results are valid (recall Fig1) & reliable (CV built-in). Even if you designed your expt w/o thinking GSS, you can apply it post-hoc! Cool, huh? Please stop using RFX. #OHBMx


OHBMX Ande Shafer 👋 1. Task MRI studies show that dorsal executive (DES; hub in dlPFC) and ventral effective systems (VES; hub in vlPFC) work in opposition to each other Goal: study the brain-behaviour relationship between resting connectivity of dlPFC and vlPFC and emotional behaviours #OHBMx


OHBMX Ande Shafer 2. 19 subjects completed Emotional Contagion Scale (sensitivity to other’s emotions) and Emotional Regulation questionnaire + 7-minute resting-state scan Seed-to-voxel connectivity maps of bilateral dlPFC and vlPFC seed regions were correlated with the behavioural scores #OHBMx


OHBMX Ande Shafer 3.Results: At rest, dlPFC (Fig 1 for connectivity maps) and vlPFC (Fig 2) are positively connected to the corresponding DES and VAS areas, respectively. Interestingly, dlPFC was also negatively connected to default mode network regions (e.g., posterior cingulate cortex) #OHBMx




OHBMX Ande Shafer 6. Patterns at rest of the key nodes of DES and VAS networks are related to how individuals interact with and handle emotional information Future directions include studying the temporal dynamics of emotion-cognition interaction through simultaneous fMRI-EEG analyses #OHBMx

Special thanks belong to our keynote tweeters: @drBreaky, Thomas Yeo, Lauri Parkkonen, Karla Loreen Miller 🇺🇦, Laura Lewis and Lucina Uddin


Missed the first #OHBMx twitter conference? You can catch up with this amazing tool created by Anibal Sólon, browse all the talks in an easier to digest format. Thanks Anibal Sólon! anibalsolon.github.io/ohbmx/ github.com/anibalsolon/oh…

Oh wow!! Check out #neuromatch2020 conference! It works so beautifully! Congratulations Kording Lab 🦖 & Dan Goodman & the whole team organising it with you! 😍🥳 Anibal Sólon Işıl Bilgin All 🧠sciency tweeps: Find the agenda below and join! It's amazing! OHBM Roberto Toro [email protected] r3rt0.bsky.social
