I love the way you overexcite

But first… Yo, a new arch dropped!

Should’ve screenshotted myself, we’re on step 6 now…

🔖 Paper link

• Using STAR genome - hierarchical NN architecture encoding

• Evolutionary algorigthm to improve the network architecture for a particular case

• Cache ⬇️90% compared to transformers and ⬇️37% compared to hybrid models (e.g. Mamba-based)

• Parameter count: ⬇️13% (infer slowly slowly slowly getting faster once I start inferring it’s really hard to stop)

Anyway, putting the bets on Liquid AI as both their liquid network and STAR are good (IMO):

So the mom and pa were right?

• Exclusion of any food group of [meat, fish, vegetables] ➡️ increase of depressive symptoms Odds Ratio, adjusted for age, sex, sociodemographics etc.:

  • 🚫 Meat → 1.37 [1.24–1.52]

  • 🚫 Fish → 1.40 [1.31–1.50]

  • 🚫 Vegetables → 1.71 [1.49–1.97]

  • 🍣 Pesco-vegetarian → 1.43 [1.19–1.72]

  • 🍳 Lacto-ovo-vegetarian → 1.36 [1.09–1.70]

• Full table:

  

• Don’t deprive yourself of more than 1 product:

• A reasonable Reddit comment: where’s ketogenic in here?

• That, and kynurenine effect (wrote on it while postCOVID bitchslapped everyone):

  • ⬆️ Kynurenic acid (⬇️ excitability + neuroprotective + decreased in depression and bipolar + antagonist at NMDA, AMPA, Kainate receptors)

  • ⬇️ Kynurenine (decreased in bipolar + increased in depression, schizophrenia and Alzheimer’s + metabolizes into potentially excitotoxic Quinolinic acid)

And, huh (don’t forget the 6%/year kidney stone risk, and time accordingly)

So, bipolar…

Lithium is modulatory. How exactly?

Been rewriting a lithium overview before, time to add some more insights.

Mechanisms of lithium, revisited

• Inhibits GSK-3b (influences metabolism and the whole nervous system cell lifecycle) → neuroprotection and normotimic effects. But you’re not here for that, right?

• ⬆️ glutamate reuptake with chronic use

• ⬇️ postsynaptic NMDA activation ← ⬇️ phosphorylation (papa ATP gives the P to tha cell, and she wan the P so bad ayy)

  

• ⬇️ excess dopamine signalling in PFC (presumably akin to what aripiprazole does?)

---

Inflammation/autoimmune and mental health

• A recent study had shown 48.1% of bipolar patients have autoimmune or allergic diseases

• Cytokines influence kynurenine pathway at large in a complex interplay:

  

• Cytokines can also influence BBB permeability → antibodies and molecular mimicry agents going in and out like it’s their home

• Different studies incl. Modabbernia et al. (2013) found that bipolar has:

  • ⬆️ IL 2,4,6, TNF-α, sIL-6R

  • Mania → ⬆️ IL-1 antagonist (IL-1RA), sIL-2R & sTNFR1

  • Depression → ⬆️ IL-10 (weird as it’s immunomodulatory)

  • ⬆️ IL-1β & kynurenic acid in spinal fluid

  • Post-mortem: ⬆️MyD88, NF-κB → IL-1 pathway activators

    

  • ⬆️ PDE4B → ⬆️ inflammation due to cAMP degradation

  • ⬆️ neutrophil/lymphocyte ratio, ⬆️ RBC/lymphocytes during mania

---

Lithium and immunity (likely downstream effects)

• ⬆️ IGF-1 expression in lithium responders. IGF-1 inhibits apoptosis → ⬆️ cell survival

• ⬆️ TNF-α & IL-4 after 8 weeks of lithium

• ⬆️ IL-1β and ⬇️ IL-6 after 6 months of lithium

Conclusion (criminally simplified)

⬆️ Inflammation → ⬆️ Involvement of kynurenine pathway → ⬆️ Quinolinic acid ⬇️ 5-HTP ⬇️ Kynurenine → (chance of) ⬆️ Excitability

PVT calcium intake/myto dysfunction and bipolar

1. Tadafumi Kato, a neuroscientist at Juntendo University in Tokyo

• “In mice, the neurons in the PVT branch off into both the amygdala, influencing negative emotions like fear, and to the nucleus accumbens, influencing positive emotions linked to reward,” says Kato.

  

• “If so, it would make sense that overexcitement of the PVT would play a central role in the extreme mood swings associated with bipolar disorder,” he says.

2. Mie Kubota-Sakashita

• ⬆️ blood cell calcium in bipolar

• 🥴 altered mitochondrial calcium uptake in bipolar

Conclusion (far-fetched)

Dysfunctional mitochondrial function in PVT (Paraventricular nucleus of the thalamus) → ⬆️ calcium uptake → bipolar

Limit overexcitability and…?

As I’ve mentioned before, overexcitability brings a lot to the table - and almost never that’s tasty stuff, mostly nasty.

Things start to look that overexcitability / inflammation / mitochondrial dysfunction in separate brain regions (especially PVT given its connectivity) can lead to a lot of trouble: bipolar, schizophrenia, ASD, ADHD etc.

So, I’d say we can expect four drug MoA patterns emerging:

• directly limiting excitability or increasing plasticity → acting on ion channels or glutamate/glycine/ampakine/kainate systems, e.g. modulators like rapastinel/apimostinel/zelquistinel. Just have a look at, say, list of investigational antidepressants

• mitochondrial enhancers → overmitochondriating the possible deficits

• targeted gene therapy once potential targets are sifted out

• anti-inflammatory or anti-kynurenine-pathway-craps treatment

Memantine and bipolar...

And again, and again

• LAMotrigine + MEMantine → ⬆️ antidepressive effects and faster onset compared to LAM + PLAcebo

• This bad boy can potentially ⬆️ kynurenic acid (the good one) production

And ADHD…

Disrupted melatonin and ADHD

• ⬆️ IL-6 in ADHD

• Melatonin → ⬇️ IL-6 → any defects in melatonin secretion would lead to an increase in inflammation among other things

---

Welcome to Teleogenic❣️

Other places I cross-post (not always) to:

• Hashnode

• Medium

• Telegram

• Twitter

• LinkedIn