After TBI, CBT Has the Best Sleep Evidence So Far

CBT has the most consistent sleep signal after traumatic brain injury

For people whose sleep changed after a traumatic brain injury (TBI), cognitive behavioral therapy has the best comparative evidence available. A 2026 systematic review and network meta-analysis by Wang et al. found that CBT significantly improved insomnia severity, sleep quality, and daytime sleepiness compared with control conditions. No pharmacologic treatment in the analysis showed superior efficacy versus control on those outcomes.

That result deserves attention because sleep medications are common in clinical practice after TBI. It does not prove that every drug fails, or that every person should stop a prescribed medication. The publication does not identify the specific drugs, doses, or treatment protocols included in its pharmacotherapy category. Still, the contrast is hard to ignore: the intervention with replicated positive results was behavioral, while the pooled drug category did not beat control.

The review included 22 randomized controlled trials with 1,299 participants who had sleep disorders following TBI—about 59 participants per trial on average. It compared treatments across the Insomnia Severity Index (ISI), Pittsburgh Sleep Quality Index (PSQI), and Epworth Sleepiness Scale (ESS). Those scales measure different things: insomnia symptoms, perceived sleep quality, and daytime sleepiness. A treatment that improves one does not automatically improve the others.

Why the result is useful—and why it is not a leaderboard

Network meta-analysis combines direct trial comparisons with indirect comparisons across a treatment network. It can help when few studies compare every intervention head-to-head. Other network meta-analyses use ranking methods such as SUCRA probabilities to compare treatments, including analyses of physical therapy for chronic ankle instability and interventions for sarcopenia. But a ranking only deserves confidence when the underlying trials are sufficiently similar and precise.

Here, the available publication summary does not report SUCRA or P-score results, confidence intervals, intervention effect sizes, or a definitive treatment hierarchy. So readers should not turn this paper into a podium ceremony with CBT wearing gold, acupuncture taking silver, and a supplement wearing a lab coat in third place. The analysis identifies patterns of statistical significance across outcomes; it does not provide enough detail to quantify the size or clinical importance of each benefit.

The trials may also have differed in TBI severity, time since injury, chronicity of sleep symptoms, pain, psychiatric symptoms, and cognitive impairment. Those differences matter. A person with chronic insomnia after a mild concussion and a person recovering from a severe TBI may have very different drivers of poor sleep. Network meta-analysis assumes populations are comparable enough for indirect comparisons; the available report cannot establish that assumption.

CBT is the only intervention supported across all three measures

CBT was backed by four randomized trials for each of the three outcomes and was the only intervention that significantly improved ISI, PSQI, and ESS scores versus control. That makes it the most replicated finding in the network, not a guarantee that every TBI survivor will respond.

In a sleep-focused form, CBT typically works on behaviors and thoughts that perpetuate insomnia: spending long periods awake in bed, irregular sleep schedules, conditioned anxiety about sleep, and unhelpful sleep rules. It may include stimulus control, sleep restriction, sleep hygiene, and cognitive techniques. Those targets make practical sense after TBI, where disrupted routines, pain, anxiety, depression, and reduced activity can all worsen sleep. The trials establish treatment effects on validated self-report scales; they do not show that CBT repairs brain injury or reverses every biological contributor to post-TBI sleep problems.

The evidence still has a blind spot. The available literature does not include a TBI-specific randomized trial showing weak or null CBT-I effects, and positive trials can be easier to find and publish than negative ones. Four trials are far better than one, but they are not the final word.

Acupuncture has modest support; the newer tools have signals, not answers

Acupuncture was significant on the ISI and PSQI in two trials, but not on the ESS. That pattern supports a modest claim: it may improve reported insomnia symptoms and subjective sleep quality in this population. It does not establish an effect on daytime sleepiness. Two trials provide more reassurance than a single experiment, yet the evidence remains thinner than the CBT evidence.

Several interventions had positive findings in only one trial. Transcranial direct current stimulation (tDCS) significantly improved ISI and PSQI scores in one study. Branched-chain amino acids (BCAAs) significantly improved ISI in one study. Hyperbaric oxygen therapy and PST each significantly improved PSQI in one study. Single-trial results can point researchers toward worthwhile questions, but they are vulnerable to chance findings, publication bias, and design quirks.

The tDCS result is especially hard to act on because the available report does not provide the stimulation target, electrode montage, current density, session length, or number of sessions. Without those details, clinicians and researchers cannot reproduce the protocol. A separate 2025 review of tDCS plus gait training in Parkinson's disease addressed motor outcomes, not sleep. It therefore cannot validate tDCS as a sleep tool for healthy people or for people without TBI.

BCAA enthusiasm should face the same reality check. The trial's dose and effect size are not reported in the available summary. A proposed explanation involves competition with tryptophan at the blood-brain barrier transporter LAT1 after TBI-related neuroinflammation. That is biologically plausible, but it remains a hypothesis rather than proof from this trial. Even if the mechanism proves correct, it depends on post-injury physiology and offers no empirical basis for using BCAAs to optimize sleep in healthy people.

Related TBI studies do not fill the sleep evidence gap

It is tempting to gather every non-drug brain intervention into one optimistic pile. The data do not permit that. A 2025 trial of transcranial photobiomodulation in mild TBI reported improvements in cognitive function, post-concussion symptoms, and PTSD symptoms. Sleep was not its primary reported outcome, so it does not establish a sleep benefit. It does, however, show why sleep research after TBI should account for cognitive and psychiatric symptoms that often travel with insomnia.

Likewise, a 2024 open pilot of brief mindfulness training in military personnel with mild TBI reported improvements in stress, pain, emotion regulation, and attention. It had no control group, and sleep was not a primary outcome. Natural recovery, expectancy effects, and regression to the mean can produce apparent improvement in an uncontrolled study. The pilot is useful for generating hypotheses, not for proving that mindfulness treats post-TBI insomnia.

Broader network meta-analyses can produce credible comparative results when their evidence base is large and well characterized. For example, a 2024 network meta-analysis of probiotic combinations for depression, anxiety, and stress included 45 trials and 4,053 participants and rated network evidence with GRADE. Wang et al.'s 22-trial, 1,299-person sleep network addresses a more specific and clinically difficult population, but its smaller trials and incomplete reporting demand more caution.

What to do with this evidence

If you have persistent sleep problems after TBI, CBT or CBT-I is the clearest evidence-based conversation to have with a clinician experienced in brain-injury care. Acupuncture may be a reasonable adjunct for people interested in it, with the understanding that its evidence is narrower and limited to insomnia symptoms and subjective sleep quality.

Medication decisions require more nuance. The review's pharmacotherapy category did not outperform control, but unnamed agents cannot be treated as interchangeable. A prescribing clinician should weigh the treatment target, potential cognitive effects, interactions, pain, depression, anxiety, and sleep-disordered breathing. The sensible response is a medication review, not an abrupt stop.

For healthy biohackers, this review offers no evidence that tDCS or BCAAs improve sleep. Their signals arose in people with TBI, each from a single trial, with missing protocol details. The evidence supports a narrower conclusion: CBT currently has the most consistent randomized-trial support for post-TBI sleep disorders, and the rest of the field needs larger, better-reported trials before it earns stronger claims.