This precisely captures one popular hypothesis that nearly all brain interventions (drugs, brain stimulation, etc..) as well as major life changes (creating a different environment, 1 month long therapeutic retreat) are basically different ways of perturbing an individual from their current sub-optimal local minima.
It is somewhat interesting to note that most interventions only roughly help people 30% of the time entirely and the rest tend to have relapses or never respond. One reason for this being that if your intervention doesn't fully address the complex system of depression pathology, then there still remain a lot of forces drawing the individual back into their depressive state.
Minor note though, I personally wouldn't try to take a particular algorithmic metaphor (simulated annealing) too seriously. For one thing I don't think the perturbations/interventions are that random, but rather tend to have insufficient coverage.
If you are interested in this topic, when I watched Ken Kendler's Oxford Lectures in 2014, it inspired me to spend some of my academic career in Psychiatry.
Good talk, thanks for sharing. I've only watched part 1 so far but the example of nicotine receptor genes functioning as pseudo-oncogenes through their effect on smoking is a really interesting example of gene -> behaviour -> environment feedback loops
This precisely captures one popular hypothesis that nearly all brain interventions (drugs, brain stimulation, etc..) as well as major life changes (creating a different environment, 1 month long therapeutic retreat) are basically different ways of perturbing an individual from their current sub-optimal local minima.
It is somewhat interesting to note that most interventions only roughly help people 30% of the time entirely and the rest tend to have relapses or never respond. One reason for this being that if your intervention doesn't fully address the complex system of depression pathology, then there still remain a lot of forces drawing the individual back into their depressive state.
Minor note though, I personally wouldn't try to take a particular algorithmic metaphor (simulated annealing) too seriously. For one thing I don't think the perturbations/interventions are that random, but rather tend to have insufficient coverage.
If you are interested in this topic, when I watched Ken Kendler's Oxford Lectures in 2014, it inspired me to spend some of my academic career in Psychiatry.
Part 1: https://www.youtube.com/watch?v=jP79G2Aozp4
Part 2: https://youtu.be/VocmxJS3s3A?si=EmTsZRAL9HJiNtSV&t=3352
Nearly 10 years later, we've only done a tiny bit to address the fundamental structural issues he discusses in part 2.
Good talk, thanks for sharing. I've only watched part 1 so far but the example of nicotine receptor genes functioning as pseudo-oncogenes through their effect on smoking is a really interesting example of gene -> behaviour -> environment feedback loops