Etiology and pathogenesis
Sleeptalking is sometimes associated with sleep loss, stress, psychopathology, or medical illness with fever. An anecdotal report exists concerning sleeptalking and pathological laughter association with clival chordoma (30). Sleep talking and delirium were reported as presenting symptoms of Hashimoto encephalopathy in a female patient with antithyroid peroxidase antibodies (TPO-Ab) in her CSF with EEG frontal slowing showing brain MRI abnormalities at the level of the commissural magna cerebri. A 3-day intravenous methylprednisolone (500 mg/day) treatment was able to promptly revert symptoms (22). Sleeptalking is a common finding in patients with Parkinson disease (44). In a study of 12 patients with Parkinson disease who screamed during sleep, sleeptalking was in several cases retrospectively identified to have preceded the onset of Parkinson disease (32). A paper reports increased frequency of nocturnal vocalization in patients with pure autonomic failure compared to idiopathic Parkinson disease and dementia with Lewy bodies, suggesting that dream enactment typical of REM sleep behavior disorder is equally shared by these disorders despite the absence of motor symptoms and dopaminergic involvement in pure autonomic failure (26). The same extent of brainstem lesions critical to REM sleep behavior disorder probably exists in all conditions, with Lewy bodies seen in pure autonomic failure, idiopathic Parkinson disease, and dementia with Lewy bodies (14).
Sleeptalking has been described in the context of IgLON5 (39) with a prevalence greater than 80%. In particular, N2 is characterized in this syndrome as poorly structured with frequent reports of sleep mumbling, vocalizations, laughing, and crying.
In this context, but also in the general population, sleeptalking is often precipitated by episodes of apneas or sleep disordered breathing. Gaig and colleagues describe simple and complex vocalizations and movements elicited by termination of apnea episodes, mostly in NREM sleep, rarely in REM (18). The condition tends to endure into adult life with persistent obstructive sleep apnea.
In an experimental protocol, Uguccioni and colleagues assessed declarative verbal learning in REM behavior disorder patients versus control subjects (40). Nighttime verbal consolidation was normal in REM behavior disorder patients unlike daytime consolidation, with incorporation of learned material with REM sleep talking in 1 out of 18 patients. Sleeptalking has also been described as part of NREM parasomniac behaviors (NPBs) in Parkinson disease and multiple system atrophy patients. More specifically elementary NREM parasomniac behaviors consisted of lingering, talking, and purposeless slow movements, whereas confusional NREM parasomniac behaviors, only typical of Parkinson disease patients, represented orienting responses often accompanied by vocalization as if calling out to someone. Most of these behaviors were triggered by respiratory events. NREM parasomniac behaviors are thought to be accounted for by primitively abnormal prefrontal and frontal networks in Parkinson disease or to be a consequence of dopaminergic agents used by these patients.
Sleeptalking has been associated with atypical sexual behavior during sleep, or sleep-related eating disorder, or both (29), during confusional arousals or sleep walking episodes.
Only in adults, it is more often associated with psychological or psychiatric abnormalities (23). The content of the sleeptalking should be taken lightly. Subjects sleeptalk in any of their dominant languages; however, immediate past events seem to influence the language used (41).
An interesting analysis and classification of sleeptalking content in terms of brain state has been published on account of Dion McGregor’s sleeptalking episodes (06). Dion McGregor (born in 1922) was the most extensive sleeptalker ever recorded. By using the Hall and van de Castle Content Scales typically applied to analyze dreams, the authors compared sleeptalking episodes from McGregor to normative male dreams (19; 21). By doing so, they found fewer female characters, but more familiar characters and friends, and less aggression, friendliness, or sex per character, but significantly more self-negativity in McGregor’s narratives, according to the Hall and van de Castle scale. As for bizarreness (Hobson’s scale), no differences were found between sleeptalking and dreams in 3 out of 6 scales, as far as discontinuity, but less incongruity of characters, objects, and actions and fewer incongruities or uncertainty of thoughts. The authors tried to explain these results as due to the differences in frontal activation in the 2 different states: sleeptalking and dreaming (33). In fact, the decreased bizarreness in sleeptalking is congruous with the known sleeptalking physiology in terms of partial reactivation of the dorsolateral prefrontal cortex (DLPC) due to increased alpha preceding the central sleeptalking episode. Furthermore, the increased activity of the subject (McGregor) in the sleeptalking episodes as a befriender or aggressor is likely related to his state (self-agency) with an EEG closer to waking, against the passivity of dreaming.
According to Alfonsi and colleagues, sleeptalking reflecting ongoing dream content may represent a unique possibility to directly access mental experience during sleep (02).
REM sleep-associated sleeptalking correlates with dream recall frequency as a function of specific neurophysiologic patterns of activation (12). A specific coherence of somniloquy with the subject’s affective tone has been reported. In this sense, rather than “acting out” dreams as in REM behavior disorder, sleeptalking could be seen as a sort of “speaking out” dreams (02).
As far as psycholinguistic processes are concerned, sleeptalking provides a unique opportunity to examine local arousals in specific linguistic circuits activated in the sleeping brain. Arnulf and colleagues showed that syntactic and prosody rules of conversation seem to be preserved during sleep even if the phonatory system appears to be inhibited, as if independent from attention-demanding processes (04). It remains to be seen whether these recruited higher cerebral areas are the same in different sleep stages or in terms of clinical relevance in disorders of arousals vs REM behavior disorder.
Sleeptalking should also be explored in relation to the hypothesis of memory activation of newly acquired information during sleep. Neuronal replay of recent memories in slow wave sleep with song replay in birds learning to sing has been demonstrated in experimental animal studies (11). Brain imaging and electroencephalographic studies in humans also support the sleep learning-dependent hypothesis (31; 34).