In multiple sclerosis, neurofeedback and biofeedback do not target the disease itself but its symptoms: memory and attention difficulties, muscle weakness, bladder and bowel dysfunction, fatigue. The evidence is thin but better than it first appeared. An older randomised trial found neurofeedback reduced fatigue and depression compared with usual care, with the gains still present two months later, and a second study found that the patients who actually learned to change their brain activity were the ones whose memory and attention improved — a link confirmed on brain imaging. A more recent trial shows an added gain when neurofeedback is combined with cognitive rehabilitation, though neurofeedback alone was less effective than cognitive rehabilitation alone. EMG biofeedback, pelvic or muscular, is the better-documented side, on very small trials. For fatigue, the one available study finds no effect of single-session respiratory biofeedback.
Updated :
What the research shows
EEG neurofeedback — standard amplitude training
(protocols unspecified) — cognition and fatigue in multiple sclerosis
Ages
Adult (18+)
Techniques
Other neurofeedback
AAPB0Not ratedNeuroLogic3Probably efficacious
Level 3 (multiple sclerosis is outside the AAPB book, so this level is entirely NeuroLogic's). Raised from 2 on pre-2022 evidence the first pass did not carry. Choobforoushzadeh 2015 randomised 24 patients with MS and primary fatigue and depression to sixteen sessions of neurofeedback or treatment as usual, with significant reductions on the Fatigue Severity Scale and the HADS depression subscale that held at two-month follow-up. Kober 2019 trained fourteen patients on SMR (12-15 Hz) over ten home-based tele-rehabilitation sessions: seven learned to modulate the rhythm and improved on long-term memory and executive function, seven did neither; Pinter 2021, in the same sample, found the responders' gains tracked increases in fractional anisotropy and in salience- and sensorimotor-network connectivity. Ayache 2021 reads the same body as scarce but preliminarily supportive. A randomised comparison against usual care, plus a within-sample responder analysis with a physiological correlate, is what level 3 describes. It is not level 4: nothing has been tested against sham or a bona fide active treatment, Motamedi 2026 still has neurofeedback as its weakest active arm, and the samples are 24 and 14.
Biofeedback
Pelvic-floor EMG biofeedback — bladder and anorectal dysfunction
Ages
Adult (18+)
Techniques
Pelvic-floor EMG
AAPB0Not ratedNeuroLogic2Possibly efficacious
Level 2. Raheem 2025: 45 patients of both sexes with overactive bladder randomised into three equal groups — biofeedback plus pelvic-floor exercise, transcranial magnetic stimulation plus exercise, or exercise alone — over six weeks. All urodynamic parameters improved within each group, but there was no significant difference between the three groups after the intervention, despite the authors concluding in favour of EMG biofeedback. Kovari 2022: individualised six-month programme (ten weeks of supervised biofeedback physiotherapy then three months of self-directed practice) in patients with anorectal dysfunction; ten of twenty patients completed the study, the St. Mark's faecal incontinence score fell from 14.00 to 9.70 then 9.30 (p = 0.005), with no significant improvement in any high-resolution anorectal manometry reading. Özkan 2026: PRISMA-ScR scoping review of 15 studies, four of them biofeedback, judged promising for symptoms and quality of life but limited by small samples, heterogeneous protocols and non-disease-specific outcomes; Sparaco 2022 reaches the same conclusion for pelvic-floor rehabilitation as a whole. A thin controlled base, with no demonstrated randomised superiority.
EMG biofeedback in motor rehabilitation — strength, functional disability and pain
Ages
Adult (18+)
Techniques
EMG biofeedback
AAPB0Not ratedNeuroLogic2Possibly efficacious
Level 2. Miri 2025: 20 patients with multiple sclerosis randomised to an EMG-biofeedback-guided isometric quadriceps programme (n = 10) or the same exercise programme alone (n = 10), five days a week for six weeks, with measurements at baseline, weeks 2, 4 and 6 and at week 12. In between-group comparisons, pain intensity, maximal isometric quadriceps strength and functional disability were significantly better in the biofeedback group at the end of week 6 (p < 0.05); the authors themselves call for caution given design constraints. Chasiotis 2023, a systematic review of six cerebellar-ataxia rehabilitation protocols (145 patients in total, three randomised trials and three pilot studies), includes visual and robotic biofeedback re-education among the techniques followed by symptom reduction after the intervention, with methodological quality rated moderate. A single very small randomised trial with between-group superiority, and no independent replication.
Respiratory and HRV biofeedback — fatigue and autonomic regulation
Level 1. Garis 2023 is the only trial in the window: 34 patients with multiple sclerosis randomised, after a cardiovagal integrity test and a fatigue-inducing vigilance task, to a biofeedback-supported deep-breathing exercise or to progressive muscle relaxation. Only patients in the progressive muscle relaxation group showed a significant drop in state fatigue, while both exercises changed autonomic activity; the vigilance task did raise short-term fatigue and reduce parasympathetic heart-rate-variability indices, and trait fatigue correlated negatively with those indices during the second half of the task. A single session, an immediate state-fatigue outcome and an active comparator: the design does not test respiratory biofeedback as it would be delivered clinically, over several weeks. No other biofeedback study of fatigue or mood in this indication across 2022-2026; this is an absence of evidence rather than evidence of absence.
In short
Clinical reading
Indication outside the AAPB 2023 book (no AAPB level). NeuroLogic Level 3 for amplitude neurofeedback in adults, raised from 2 on pre-2022 evidence: Choobforoushzadeh 2015 (randomised, n = 24, sixteen sessions vs treatment as usual, significant on fatigue and depression at post and at two months), Kober 2019 (n = 14, ten home SMR sessions, seven responders improving on memory and executive function) with Pinter 2021 showing responders' gains tracking fractional-anisotropy and network-connectivity increases in the same sample, and Ayache 2021 as narrative support. Motamedi 2026 (n = 30, combined arm superior, neurofeedback alone inferior to cognitive rehabilitation) and the uncontrolled review literature (Vilou 2023) remain as before; the multiple-sclerosis studies were excluded from Ribeiro 2023's SMR meta-analysis for want of a control group. No sham comparison exists in the indication, which is what holds the level at 3. Pelvic-floor EMG biofeedback 2 (Raheem 2025 with no between-group difference; Kovari 2022; Özkan 2026 scoping review), motor EMG biofeedback 2 (Miri 2025, n = 20), respiratory and HRV biofeedback for fatigue 1 (Garis 2023, single session, negative).
Protocols
Neurofeedback: protocols unspecified in the available literature, clinic-delivered twice weekly and often paired with cognitive rehabilitation; twelve sessions in the only randomised trial, below the 20-40 usual for amplitude work. Biofeedback: pelvic-floor EMG alongside pelvic-floor exercise over six weeks to six months, with home practice; surface EMG guiding isometric quadriceps strengthening five days a week for six weeks; biofeedback-supported slow breathing for fatigue.
Limits
Very small samples (n = 20 to 45), a single randomised neurofeedback trial whose protocol is not reported, no sham-controlled trial, blinding rarely stated, follow-up at best to one month. Fatigue and quality-of-life outcomes are not disease-specific in the pelvic-floor biofeedback studies. An adult indication: no paediatric literature, and paediatric multiple sclerosis is not covered. Disability progression and imaging parameters are never outcome measures.
Study base
2022-2026 base: 5 publications indexed in the archive (1 neurofeedback, 4 biofeedback), together with reviews found in the search (Vilou 2023, Bossa 2022, Sparaco 2022, Chasiotis 2023) and Ribeiro 2023's SMR meta-analysis, which excluded the multiple-sclerosis studies.
Brendan's perspective
Level 3, and it is ours, since multiple sclerosis is outside the AAPB book. This is a raise from 2, and it comes from going back before the AAPB window rather than from anything new. Choobforoushzadeh 2015 randomised twenty-four patients with MS to sixteen sessions of neurofeedback or treatment as usual and moved both fatigue and depression, with the gains still present at two months. Kober 2019 trained fourteen patients on SMR at home: seven learned to modulate the rhythm and improved on memory and executive function, seven did neither — and Pinter 2021 showed that the responders' gains tracked measurable changes in connectivity and white-matter microstructure. That last pairing is the part I find persuasive, because it ties the clinical change to the thing that was trained rather than to the fact of being trained. Ayache 2021 reads the same literature as scarce but promising, and I agree with both halves. It is not a 4: nothing has been tested against sham or a bona fide active treatment, and Motamedi 2026 — twelve sessions, ten per arm, sites and bands unreported — still has neurofeedback as its weakest arm, which is not a fair test of amplitude training as I would deliver it. I would offer it as an adjunct to cognitive rehabilitation, with autonomic work alongside, and at 20 to 40 sessions rather than the ten to sixteen these studies ran. The better-documented half is EMG biofeedback inside physiotherapy — Miri 2025 for quadriceps strength, pelvic-floor training for bladder symptoms, where Raheem 2025 found no difference between its three groups. What a clinician should say: nothing here has been tested against sham, no outcome touches disability, and follow-up stops at a month.
Motamedi et al. (2026) Combined cognitive rehabilitation and neurofeedback therapy: effects on executive function and cognitive flexibility in people with multiple sclerosis doi:10.1007/s10072-025-08644-3
Vilou et al. (2023) EEG-Neurofeedback as a Potential Therapeutic Approach for Cognitive Deficits in Patients with Dementia, Multiple Sclerosis, Stroke and Traumatic Brain Injury doi:10.3390/life13020365
Ribeiro et al. (2023) Clinical applications of neurofeedback based on sensorimotor rhythm: a systematic review and meta-analysis doi:10.3389/fnins.2023.1195066
Raheem et al. (2025) Comparative study on the Impact of Transcranial Magnetic stimulation and Bio-feedback on overactive bladder in multiple sclerosis patients: a Randomized Clinical Trial doi:10.1007/s10072-024-07788-y
Kovari et al. (2022) Anorectal dysfunction in multiple sclerosis patients: A pilot study on the effect of an individualized rehabilitation approach doi:10.3233/NRE-210226
Özkan et al. (2026) Non-pharmacological interventions for fecal incontinence in people with multiple sclerosis: A scoping review doi:10.1016/j.msard.2025.106864
Miri et al. (2025) Assessing the effectiveness of EMG-biofeedback on an isometric strengthening program in individuals with multiple sclerosis doi:10.1016/j.msard.2024.106193
Garis et al. (2023) Comparing two relaxation procedures to ease fatigue in multiple sclerosis: a single-blind randomized controlled trial doi:10.1007/s10072-023-07042-x
Choobforoushzadeh et al. 2015 (2015) Effect of neurofeedback training on depression and fatigue in patients with multiple sclerosis doi:10.1007/s10484-014-9267-4
Kober et al. 2019 (2019) Self-regulation of brain activity and its effect on cognitive function in patients with multiple sclerosis doi:10.1016/j.clinph.2019.08.025
Pinter et al. 2021 (2021) MRI correlates of cognitive improvement after home-based EEG neurofeedback training in patients with multiple sclerosis: a pilot study doi:10.1007/s00415-021-10530-9
Ayache et al. 2021 (2021) Neurofeedback therapy for the management of multiple sclerosis symptoms: current knowledge and future perspectives doi:10.31083/j.jin2003079
