Thread Reader
Miriam Klein-Flügge 🧠🇪🇺🇺🇦

Miriam Klein-Flügge 🧠🇪🇺🇺🇦
@MKFlugge

Sep 22, 2022
20 tweets
Twitter

I'm extremely happy to share our new study now out in @NatureHumanBehaviour - with @Daria Jensen, Yu Takagi, Luke Priestley, @Lennart Verhagen @Stephen Smith and Matthew Rushworth: nature.com/articles/s4156…

We wanted to understand which brain circuits reflect specific aspects of people’s mental well-being. We looked at data from nearly 500 people from the @HCP News project for which we had both functional connectivity and mental health measures.
Two aspects of our approach were important and different from previous work. First, we looked at brain connectivity with the amygdala – a region well known for its importance in emotion and reward processing – but at a finer spatial scale than previously done.
We parcellated the amygdala into 7 nuclei and looked at functional connectivity with these specific nuclei and other subcortical/cortical regions of interest. This parcellation was robust, replicable, relatively symmetrical and in line with post-mortem and NHP work.
The amygdala parcellation is available here: osf.io/egm2r/
Second, we looked at mental well-being in healthy people on a continuous scale and along several dimensions. Instead of using broad summary scores, we separately captured people’s social well-being, emotional well-being, sleep quality and problems with anger/rejection.
In the critical analysis step, we hypothesized that with our improved level of detail for both brain connectivity and individuals’ well-being, it may be possible to identify precise and dissociable brain networks that relate to these distinct aspects of mental health.
Indeed, this is what we found. The brain connections that matter most for discerning whether someone is struggling with sleep problems, for example, look very different from those that carry information about someone’s social well-being (i.e., having friends + feeling connected)
And the nature of the identified brain networks differs too, for example negative emotions relate most strongly to connectivity in older subcortical circuits, while amygdala connectivity with both cortical and subcortical circuits reflects people's social well-being.
We replicated all our results in the held-out 7T HCP dataset which suggests these results are robust, despite known problems with low signal-to-noise ratios in the amygdala; please see our Methods for more details on what steps we took to improve the data quality in subcortex.
The held-out predictions were better when using nuclei-specific connectivity as opposed to whole amygdala connectivity, and when using dimensional mental health markers as opposed to summary scores such as the total depression score
We think it is important to understand how changes in the brain and changes in well-being relate. We suggest to study the affected behaviours and the involved brain networks at a finer scale than typically done – a scale that might better match the brain’s functional organization
These findings will not change how we treat psychiatric disorders, but I believe they do give hope. In the future, instead of choosing treatments by trial and error, it may be possible to target treatments to the brain circuits most relevant for an individual’s key symptoms.
This possibility is becoming more tangible with current progress in non-invasive deep brain stimulation methods (e.g., with ultrasound). I believe an improved understanding of the neural circuitry related to specific mental health dimensions is an important step in this direction
Thank you so much to my amazing collaborators, the @HCP News team and to the anonymous reviewers who made this paper better.
The personal twist: I will be honest - this project was a long haul. It took ~8 years to complete, it was tricky for many reasons and revisions during Covid/home-schooling were very slow. But we very much hope that you will find it useful, and we look forward to your feedback!
We cannot publish the data itself because it is part of HCP (but you can apply for access with them directly here: humanconnectome.org). But everything else including the parcellation and the code is available here: osf.io/egm2r/
Thanks for reading until here… Feel free to share and/or comment!
Access to the full-text version of this article should be possible via this link: rdcu.be/cV7tA
Miriam Klein-Flügge 🧠🇪🇺🇺🇦
cognitive neuroscientist, associate prof, mum, decision making, ultrasound neuromodulation & mental health @OxfordWIN @OxExpPsy @OxPsychiatry (she/her)
Follow on Twitter
Missing some tweets in this thread? Or failed to load images or videos? You can try to .