3D Virtual Reality Performance Metrics as a Future Fatigue Biomarker in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS), 2026, Ladek et al

[forestglip]

3D Virtual Reality Performance Metrics as a Future Fatigue Biomarker in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS)

Spoiler: Authors

Ladek, Anja-Maria; Priebe, Leonie; Harrer, Thomas; Harrer, Ellen; Michelson, Georg; Knauer, Thomas S.; Dias-Nunes, Diogo X.; Mardin, Christian Y.; Bergua, Antonio; Hohberger, Bettina

Background
Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) is a debilitating disorder, characterized by symptoms such as post-exertional malaise (PEM) and cognitive impairments. This study assessed reaction time (RT) metrics in three-dimensional (3D) visual tasks with the aim of objectively quantifying the cognitive impairments in ME/CFS patients compared to controls.

Methods
A total of 120 participants (60 ME/CFS patients and 60 controls) were recruited at the Department of Ophthalmology, Universität of Erlangen-Nürnberg. RT was assessed using a virtual reality–oculomotor test system, presenting 3D stimuli at three disparity levels (275″, 550″, and 1100″) within three gaming repetitions (R1, R2, and R3). Mixed-effects models were used to evaluate group differences, with age and gender as covariates.

Pairwise contrasts were calculated to assess changes across repetitions. Fatigue self-assessments were recorded by validated questionnaires, (FACIT Fatigue Scale, Chalder Fatigue Scale, Bell Score and Health Assessment Questionnaire), and their correlation with RT metrics was portrayed using a Spearman correlation matrix.

Results
Estimated means (EM-means) for RT were significantly prolonged in ME/CFS patients compared to controls at disparity 275″ (1969 ms vs. 1384 ms; p = 0.0001), 550″ (1409 vs. 1071 ms; p = 0.0012) and 1100″ (1126 ms vs. 891 ms; p = 0.00223). Age was a significant covariate (p < 0.001), while gender showed no effect.

Both groups demonstrated improvements in RT over repetitions; however, ME/CFS patients showed a significantly lower improvement compared to controls, reaching significance in R3 (p = 0.0042). RT metrics did not correlate with patients’ self-assessment scores.

Conclusions
ME/CFS patients showed consistently slower RTs compared to controls, particularly in later, easier gaming repetitions, potentially reflecting the impact of fatigue.

Web | DOI | PDF | Biomedicines | Open Access

 

[Utsikt]

They really go after the questionnaires when discussing the lack of correlation with the objective results:

We had hoped to be able to show correlation between the RT metrics in VR-OTS performance and the self-assessment questionnaires (patient-reported outcome measures, PROMs). Neither anamnestic reports nor PROMs correlated with any RT outcome metrics. The greatest correlation was achieved between the Chalder Fatigue Scale and the RT at 275″ disparity. Other RT metrics showed singular correlations with the Chalder Fatigue Scale without showing a clear pattern.

Rather than supposing that the predictive value of VR-OTS for ME/CFS severity is low, we assume that although these questionnaires are commonly used in diagnostics of ME/CFS and help in observing changes within a person, they are subjective.

Difficulties comparing greater populations have been documented in other studies as well [36,37], particularly when the patient’s own interpretation of each question is allowed in self-assessment. This is supported by the literature, as ceiling effects and bad internal consistency [36,37] as well as poor prediction across subscales has led to PROMs being criticized even beyond scales used in ME/CFS [38,39].

Comparison of questionnaire assessments and sensor measurements for time spent upright showed not only overestimation of actual time spent upright by the ME/CFS cohort but also demonstrated the lack of correlation between PROMs and time spent upright in ME/CFS cohorts [40].

In line with this, Stussman et al. [13] recently showed that qualitative interviews are more sensitive than questionnaires when quantifying PEM changes.

 

[forestglip]

ME Research UK: "Could reaction times to 3D virtual reality tasks act as a biomarker for fatigue in ME/CFS?"

'Although this was a relatively small study, the results showed that when compared with those of healthy controls, reaction times of those with ME/CFS were significantly slower. It was also observed that although reaction times generally improved with practice, those of the ME/CFS group remained slower than those of healthy controls.​

Interestingly in this study, there was no significant association between reaction time and symptom severity. However, this finding may be explained by a lack of diversity in severity levels within the ME/CFS group, i.e., if all ME/CFS participants had mild symptoms.​

While there are limited conclusions that can be drawn from this study, the team state that “A larger-scale study has recently begun with the aim to correlate biomarkers with performance. This identification and validation of objective biomarkers may help in diagnosis, disease awareness and understanding.“​

It is also worth highlighting that these findings support the hypothesis of a ME Research UK–funded PhD study by Anosha Altaf, which proposes that reduced nerve responses in ME/CFS are associated with slower, less accurate responses to visual stimuli.'​

 

[ScoutB]

reduced nerve responses in ME/CFS are associated with slower, less accurate responses to visual stimuli.

Interesting.

Since AMPA receptors came up (from discussion of SHISA6 here), I've had this open in a tab hoping to make sense of it when less foggy: Neocortex saves energy by reducing coding precision during food scarcity.

From what I understand, the idea is that the neurons being studied go into an energy saving, but more error prone, state when they sense the mouse has less energy (via leptin, in this case).

Decoding performance reflects how well neuronal population activity in response to distinct visual stimuli can be discriminated. We found that scenes from very different environments (outdoor versus home cage) could be decoded equally well in both control and food-restricted groups; however, decoding of similar scenes taken from the same environment was significantly impaired in the food-restricted group. These findings indicate that under food restriction, population coding of visual information was largely preserved, but coding precision was reduced.

It would sort of track with the general picture of "feeling awful, brain not working properly, yet no obvious damage", if some ME/CFS neurons turned out to be in some kind of low-energy, bad-at-doing-neuron-things state..