Hypothalamus Connectivity in Adolescent Myalgic Encephalomyelitis/Chronic Fatigue Syndrome, 2024, Byrne et al.

SNT Gatchaman

Senior Member (Voting Rights)
Staff member
Hypothalamus Connectivity in Adolescent Myalgic Encephalomyelitis/Chronic Fatigue Syndrome
Hollie Byrne; Sarah J. Knight; Elisha K. Josev; Adam Scheinberg; Richard Beare; Joseph Y. M. Yang; Stuart Oldham; Katherine Rowe; Marc L. Seal

Adolescent Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is a disabling illness of unknown etiology. Increasing evidence suggests hypothalamic involvement in ME/CFS pathophysiology, which has rarely been explored using magnetic resonance imaging (MRI) in the condition. This work aimed to use MRI to examine hypothalamus connectivity in adolescents with ME/CFS and explore how this relates to fatigue severity and illness duration.

25 adolescents with ME/CFS and 23 healthy controls completed a neuroimaging protocol consisting of structural and multishell diffusion-weighted imaging sequences, in addition to the PedsQL Multidimensional Fatigue Scale to assess fatigue severity. Information about illness duration was acquired at diagnosis. Preprocessing and streamlines tractography was performed using QSIPrep combined with a custom parcellation scheme to create structural networks. The number (degree) and weight (strength) of connections between lateralized hypothalamus regions and cortical and subcortical nodes were extracted, and relationships between connectivity measures, fatigue severity, and illness duration were performed using Bayesian regression models.

We observed weak-to-moderate evidence of increased degree, but not strength, of connections from the bilateral anterior-inferior (left: pd [%] = 99.18, median [95% CI] = −22.68[−40.96 to 4.45]; right: pd [%] = 99.86, median [95% CI] = −23.35[−38.47 to 8.20]), left anterior-superior (pd [%] = 99.33, median [95% CI] = −18.83[−33.45 to 4.07]) and total left hypothalamus (pd [%] = 99.44, median [95% CI] = −47.18 [−83.74 to 11.03]) in the ME/CFS group compared with controls. Conversely, bilateral posterior hypothalamus degree decreased with increasing ME/CFS illness duration (left: pd [%] = 98.13, median [95% CI]: −0.47[−0.89 to 0.03]; right: pd [%] = 98.50, median [95% CI]:- 0.43[−0.82 to 0.05]). Finally, a weak relationship between right intermediate hypothalamus connectivity strength and fatigue severity was identified in the ME/CFS group (pd [%] = 99.35, median [95% CI] = −0.28[−0.51 to 0.06]), which was absent in controls.

These findings suggest changes in hypothalamus connectivity may occur in adolescents with ME/CFS, warranting further investigation.

Link | PDF (Journal of Neuroscience Research) [Open Access]
 
Brain imaging results are all over the place across various studies, as noted in the introduction of this paper.

It's not clear to me how much of the analysis was automated and how much involved manual work that might be subject to bias. There was no statement about the people analysing the images being blinded to the ME/CFS status of the participant.

The sample size was decent for a brain imaging study, but it's still small and there were a lot of different things measured.
Thus, conceptually the likelihood of each tested model is treated independently from one another, reducing the need to correct for multiple comparisons (Sjölander and Vansteelandt 2019).
They used a statistical method that they say does not require correction for multiple comparisons. However, I think the fact remains, that if you measure a lot of different things in two groups of participants, it's not surprising to find some significant differences. The approach to recruitment creates scope for bias:
Briefly, a convenience sample of 48 participants (25 ME/CFS and 23 healthy controls) were recruited for the study between 2013 and 2016, with inclusion criteria as follows: (a) adolescents aged 13–18 years diagnosed with ME/CFS who were recruited through a tertiary hospital specialized ME/CFS Clinic using the Canadian Consensus Criteria adapted for pediatrics (illness duration ≥ 3 months) (Jason et al. 2006), and; (b) healthy adolescent controls aged 13–18 years at the time of recruitment with no history of ME/CFS or other chronic illnesses, recruited through advertising posters around the hospital.
 
3.4 Differences in Hypothalamus Connectivity Between ME/CFS and Controls
When controlling for sex, the Bayesian linear models provided evidence to suggest group status did not predict hypothalamus connectivity strength from total or hypothalamus subregions (all BF10 values ≤ 0.16), suggesting connectivity strength from the hypothalamus was relatively comparable between groups.

So strength of connectivity wasn't different.

Reported confidence intervals for relationships of connectivity degree (as opposed to strength, whatever 'degree' means) straddle either side of zero. So, I don't think they were really different either. They use the phrase 'anecdotal evidence' quite a bit. I think that means a trend that isn't statistically significant?

In the discussion, they report
The primary finding from this study is that the ME/CFS group showed evidence of increased connectivity, as measured by degree, from the bilateral anterior hypothalamus and total left hypothalamus compared with healthy controls.

As previously discussed, the anterior hypothalamic network primarily contributes to neuroendocrine, thermoregulatory and sleep processes (Burdakov and Peleg-Raibstein 2020; Chao et al. 2022). Connections from the right anterior-inferior hypothalamus, where we observed the greatest strength of evidence for differences in degree between groups, contains nuclei centrally involved in regulating circadian rhythm (suprachiasmatic nucleus), osmotic and cardiovascular processes (supraoptic nucleus), homeostasis of metabolic or reproductive processes (arcuate nucleus) and thermoregulation and satiety (anterior hypothalamic and ventromedial nuclei) (Spindler, Özyurt, and Thiel 2020). Some of these regulatory functions are known to be disturbed in ME/CFS (i.e., impaired thermoregulation, circadian rhythm, metabolism, satiety and cardiovascular disturbances; Bateman et al. 2021), however, these principally form secondary symptoms and are typically heterogenous across the population.
While the contribution of individual hypothalamic nuclei to ME/CFS symptomatology cannot currently be determined, in healthy populations, there is a body of evidence to suggest changes in anterior hypothalamic connectivity is associated with sleep disturbances

I don't think there is anything much in this study. The paper notes:
Finally, comprehensive information surrounding sleep and/or secondary symptoms (such as the type and severity of autonomic dysfunction) were not acquired during this study, which would enable a greater understanding of the contribution of symptom heterogeneity to ME/CFS pathophysiology. Acquiring such data will be vital for building a comprehensive picture of ME/CFS pathophysiology in future and bridging the gap between the conflicting imaging findings that currently exist in the literature.
 
It's not clear to me how much of the analysis was automated and how much involved manual work that might be subject to bias. There was no statement about the people analysing the images being blinded to the ME/CFS status of the participant.

I think it was automated but manually checked. The checking could be biased.

Visual inspection by a single reviewer (HB) [lead author] of the hypothalamus subregions in MNI152 1 × 1 × 1 mm space determined the ROI's were accurately defined.
 
3.4 Differences in Hypothalamus Connectivity Between ME/CFS and Controls


So strength of connectivity wasn't different.

Reported confidence intervals for relationships of connectivity degree (as opposed to strength, whatever 'degree' means) straddle either side of zero. So, I don't think they were really different either. They use the phrase 'anecdotal evidence' quite a bit. I think that means a trend that isn't statistically significant?

In the discussion, they report





I don't think there is anything much in this study. The paper notes:


Some of these regulatory functions are known to be disturbed in ME/CFS (i.e., impaired thermoregulation, circadian rhythm, metabolism, satiety and cardiovascular disturbances; Bateman et al. 2021), however, these principally form secondary symptoms and are typically heterogenous across the population.

I get that these might be heterogenous, particularly circadian rhythm, from talking to other pwme - but don't know whether that is different in adolescents? SO what do they mean by 'the population' and who suggested that and when, because we all know about all the past stuff by certain types - there's a bit of cherry-picking, reframing and perceptual preference that seems to go on.

And given how much things have moved on even just over recent years would we call these secondary symptoms? particularly when one of them is just the word 'metabolism' ie they are very general categories of symptom and not specific symptoms?

I know maybe they've been written this way because they are actually listing what those parts of the hypothalamus do, and noting a 'match process' between that and ME/CFS at whatever level. But who knows if it wasn't about the hypothalamus that these broad categories of area don't for some cover the lack of function of the body that contributes to their primary symptoms? ie exhaustion?
 
"2.2 Participants
The full procedure for participant recruitment and data collection in this study has been described elsewhere (Josev et al. 2020, 2021). Briefly, a convenience sample of 48 participants (25 ME/CFS and 23 healthy controls) were recruited for the study between 2013 and 2016, with inclusion criteria as follows: (a) adolescents aged 13–18 years diagnosed with ME/CFS who were recruited through a tertiary hospital specialized ME/CFS Clinic using the Canadian Consensus Criteria adapted for pediatrics (illness duration ≥ 3 months) (Jason et al. 2006), and; (b) healthy adolescent controls aged 13–18 years at the time of recruitment with no history of ME/CFS or other chronic illnesses, recruited through advertising posters around the hospital."
 
Could they simply be measuring common "I feel lousy, life sucks" brain patterns? Effect, rather than cause, and possibly common in other unpleasant diseases?

The problem with functional MRI is that you get to the point where they may be measuring someone thinking 'I wonder if they will find something out about my illness'. The controls, in contrast may be thinking 'this is an easy way to spend a Tuesday morning feeling virtuous'.

I don't have a feel for the biology of what they are measuring - whether it indicates new connections or just what is being thought, as you say.
 
It's not functional MRI in the way we usually see - looking at neurovascular coupling and what bits of brain light up in relation to each other. This is typically in a task-based paradigm, rather than resting state. Ie look at this card saying "green" printed in red ink. What they're doing here is diffusion tensor imaging, looking at white matter tracts, to see what is connected to what structurally.

Introduction refs removed said:
In current neuroimaging practice, DWI allows anatomical connectivity within the brain to be mapped in vivo. This method has been used extensively to develop comprehensive macroscale maps of neural connections (referred to as connectomes), which model the brain as a complex network of ‘nodes’ (brain regions) and ‘edges’ (white matter pathways connecting brain regions). Mathematical approaches such as graph theory can then be applied to describe the overall white matter organization in the brain. While graph theory contains multiple descriptive measures of network properties, most inherently relate to two simple quantities: nodal degree, describing the number of connections; and nodal strength, reflecting the weight of connections. Collectively, these two measures provide information on the variability of local connectivity of different nodes in the brain. While functional information is not contained within structural connectomes, several studies suggest these networks constrain functional integration between regions. Thus, this approach may provide useful insight into how the structural organization of hypothalamus white matter connectivity may be affected in adolescents with ME/CFS.

Earlier they described the idea to see what the effect of disease duration might be.

While white matter connectivity has rarely been studied in ME/CFS, diffusion-weighted imaging (DWI) studies examining white matter microstructure in this condition have produced heterogeneous results. […] In accordance with the existing literature on autonomic involvement in ME/CFS pathophysiology, one study by Thapaliya et al. (2021) identified decreased axial and mean diffusivity in the descending corticocerebellar tract (connecting the midbrain and pons), in addition to increased transverse diffusivity in the medulla, in adults with ME/ CFS compared to controls. However, recent work from our team did not identify such changes in the adolescent ME/CFS population.

Based on existing evidence that suggests changes in white matter microstructure may be associated with illness duration in adults with ME/CFS, the discrepancy between these findings may be attributed to the influence of older age and/or longer disease duration that are not yet established in the adolescent cohort. This suggests that, compared to whole-brain analyses, hypothesis-driven region-of-interest DWI studies may provide greater utility when examining the adolescent ME/CFS population.
 
to see what is connected to what structurally.

Thanks, yes that makes sense.
I find it hard to believe that having an illness like ME/CFS would alter connections of the hypothalamus very much. I may be wrong but it is hard to see how that would work.

What might be more interesting would be the possibility that if your hypothalamus develops in a certain way you are at risk of ME/CFS making use of its signalling loops. On the model of stochastic neural development with selection that Edelman called neural Darwinism that might be feasible although I have always seen the stochastic aspect as fine grain.

Maybe someone needs to set up a unit that collects pairs f twins with either one or both having ME/CFS. A lot could be learnt from that.

Nevertheless, I agree with others that the results here are not exactly sore thumb level.
 
and there is the noise from MRIs vs pwme - I've no idea how long these scans would have been for, or if they looked at the method in detail enough to know whether eg the tasks are genuinely specific/clean enough to only the hypothalamus bits that potential increased load on the executive function from dealing with the equivalent of a migraine/sensory could impact

I'm a bit intrigued because really surely the only way you can actually conclude something is related to sleep, circadian rhythm, appetite and so on would be things very directly related to that. Thermoregulation I can see how you might try that more.

Maybe I'm lacking ingenuity in working out what task they think can proxy these things. Because surely otherwise a task actually only definitely really tells you the pattern/connection seen for that task?
 
Back
Top Bottom