By Alan Z. Segal, MD

Associate Professor of Clinical Neurology, Weill Cornell Medical College

Dr. Segal reports no financial relationships relevant to this field of study.

SYNOPSIS: Patients with traumatic brain injuries need longer sleep times to heal the injured brain, and persistent pleiosomnia at 18 months implies that ongoing abnormalities are producing an increased need for sleep.

SOURCE: Imbach LL, Buchele F, Valko PO, et al. Sleep-wake disorders persist 18 months after traumatic brain injury but remain underrecognized. Neurology 2016; Epub April 27, 2016.

Brain injury, no matter the mechanism — traumatic, degenerative, ischemic, or demyelinating — profoundly affects function in the waking state. It is increasingly recognized, however, that sleep, a state in which the brain spends approximately 30% of its time, is also exquisitely sensitive to brain injury. Sleep is driven by the hypothalamus, but its effects are widespread; the hippocampus allows for consolidation of declarative memories, the basal ganglia facilitate muscle relaxation in REM sleep, and the calcarine cortex creates visual imagery while dreaming in REM. Sleep is physiologically necessary, particularly because it promotes the so-called “glymphatic” system within the brain, allowing toxic metabolic waste products accumulated during the waking state to be washed away.

Sleep is a delicate balance between two functions — a homeostatic drive to sleep (Process H), which builds throughout the day and is discharged as we sleep, and a circadian rhythm (Process C), which follows a sinusoidal pattern. Peak circadian alerting influence occurs in the late afternoon (e.g., 4 p.m.) and has its nadir about 12 hours later. Process C allows us to stay awake later in the day, even as Process H builds, and then Process C prevents us from waking up too early, as Process H dissipates. The strength of this intrinsic system may explain why, for example, hunter-gatherer communities, who may or may not have access to electric lights after dark, still have very consistent sleep patterns.

In prior work by these authors, traumatic brain injury (TBI), whether mild, moderate, or severe, was shown to profoundly affect sleep. This was manifested both as an increase in “sleep pressure” (measured by a shortened sleep onset latency) and an increased overall sleep need (known as pleiosomnia). The current study explores the same cohort, initially studied at 6 months, now 18 months post TBI.

Imbach et al initially studied 60 patients with TBI, classified by their initial Glasgow Coma Scale (GCS), of which 31 were still available for follow up. There were 21 mild patients (GCS 13-15), two moderate patients (GCS 9-12), and eight severe patients (GCS 3-8). Total sleep times (as measured by polysomnography) were significantly greater among TBI patients — over seven hours, compared with 6.5 hours for controls (P < 0.005). More extended home monitoring of sleep over a two-week period (as measured by movement sensing actigraphy) showed that TBI patients slept eight hours per night, compared with seven hours for controls (P < 0.00005). This was important because it proved that the effects seen during a single polysomnography continued and was validated on an extended nightly basis. TBI patients tended to fall asleep quickly, as measured by a Sleep Onset Latency (SOL) of seven minutes compared with 12 minutes for controls. Thus, TBI patients exceeded the typical cutoff for excessive sleepiness, which is defined as a SOL of less than eight minutes during a Multiple Sleep Latency Test.

Interestingly, all other measurements relating to the quality of sleep were equal. This included time in deep, slow-wave (Stage 3) sleep (measured as “delta power”) and time in all other sleep stages, most importantly, REM sleep. A computerized sleep architecture algorithm to define sleep fragmentation was identical between TBI patients and controls. Subjective ratings of daytime sleepiness, such as the Epworth sleepiness scale, were equal. Importantly, these data indicate that the problem was not the quality of sleep in TBI patients, but rather only the quantity.

In the authors’ initial study, patients with an underlying intracerebral hemorrhage (ICH) as part of their TBI showed greater sleep demands than patients without ICH. Overall, TBI severity also correlated with sleep need. This effect was lost at 18 months. It is possible that sleep, functioning as an endogenous repair mechanism required by the healing brain, played more of an influence at six months compared to 18 months. Persistent pleiosomnia at 18 months implies that there may be ongoing abnormalities in the brain that are producing an increased need for sleep. Since the 18-month data were comprised primarily of patients with mild TBI, it suggests that abnormalities such as diffuse axonal injury, or perhaps even early amyloid plaque deposition, in these patients may be much more than expected.


These data are indicative of a phenomenon rarely seen in clinical medicine. That is, a situation in which the patients’ subjective complaints actually may underestimate their objectively measurable problems. As the authors noted, TBI patients may have a component of “sleep state misperception” (also known as “paradoxical insomnia”). These patients spend longer periods of time in bed, and may complain that they are awake for much of it. In fact, when studied by objective measures, such as polysomnography or actigraphy, it is found that these patients are sleeping much more than they really think. Mild TBI patients, with post-concussive syndrome — non-specific complaints such as low-grade dizziness or mental fogginess — simply may not be sleeping enough. Given the hazards associated with excessive daytime sleepiness, such misperception may have serious public health consequences. As the authors noted, “normal” scores on subjective indices such as the Epworth (with such questions as “how likely are you to doze off stopped in traffic?”) may belie a much more insidious problem. Although there may be confounding factors, there is little doubt that in TBI, longer sleep times are needed, either to heal the injured brain or at least as a consequence of persistent CNS dysfunction.