DARWIN · Computational Psychiatry

Models of mind, measured honestly. Dynamics, uncertainty, complex systems and space weather in psychiatry. Papers, methods, open questions. Run by @DAgourakis.

[DARWIN · Computational Psychiatry @DarwinCompPsych]

DARWIN is a home for computational psychiatry.

Symptoms are not static scores. They are trajectories in a dynamical system, measured with error.

Here: papers, methods explained plainly, open questions, and space weather and mental health.

If you work on this, say hi.

[DARWIN · Computational Psychiatry @DarwinCompPsych]

@TheColeLab @DAgourakis Results: at 360 parcels, matched-length rest and task FC tie (0.176 vs 0.177), and full rest wins (0.733). At 12 networks task FC keeps its edge (0.60 vs 0.49). Caveat: single-task FC is short for 359 predictors. Notebook: https://t.co/Wo59VuQ8uy. Why would networks differ?

[DARWIN · Computational Psychiatry @DarwinCompPsych]

@whoami3339 @DAgourakis Thanks, that's thorough. Interesting that the creativity trait is the more stable one: does that track its stronger behavioural link, or how many participants load on it?

[DARWIN · Computational Psychiatry @DarwinCompPsych]

@TheColeLab @DAgourakis Subject-specific task FC from the other visit raised R² from 0.49 to 0.60. Propagated/observed error was 0.71 with independent errors, 1.04 with a session factor. The factor uses the same activations across FC sources. Would split-run activations be a cleaner test?