The neurocognitive study of sleep mentation began in 1953 when Eugene Aserinsky, in the first all-night electroencephalograph (EEG) study of sleep, observed periodic intervals of rapid eye movements (REMs) associated with reports of long, visually vivid, and sometimes bizarre dreams. REMs occur during periods of EEG so similar to waking that European scientists called it "paradoxical" sleep. REM periods make up about 22% of sleep time. Relative to REM reports, NREM reports are shorter and more thought like. Although REM reports have more bizarre elements, when corrected for report length, REM and NREM reports are equally bizarre.

This volume describes how the conceptual and technical sophistication of contemporary cognitive and neuroscientific fields has enhanced the neurocognitive understanding of dreaming sleep. Because it is the only naturally-occurring state in which the active brain produces elaborate cognitive processes in the absence of sensory input, the study of dreaming offers a unique cognitive and neurophysiological view of the production of higher cognitive processes. The theory and research included is driven by the search for the most direct relationships linking the neurophysiological characteristics of sleepers to their concurrent cognitive experiences. The search is organized around three sets of theoretical models and the three classes of neurocognitive relationships upon which they are based. The contributions to this volume demonstrate that the field has begun to move in new directions opened up by the rapid advances in contemporary cognitive science, neuropsychology, and neurophysiology.

Research in animals has demonstrated that patterns of neural activity first seen during waking experience are later “replayed” during sleep, in hippocampal and cortical networks. The characteristics of memory reactivation during human sleep, however, have not yet been fully described. Meanwhile, the possible relationship of dreaming to this “replay” of memories in the sleeping brain is entirely unknown. In the present study, we induced hippocampus-dependent memory retrieval during human sleep using a “trace conditioning” procedure. Prior to sleep, subjects underwent either trace (hippocampus-dependent) or delay (hippocampus-independent) auditory fear conditioning. Conditioned stimuli were then presented to subjects during non-REM sleep. Both delay-conditioned and trace-conditioned subjects exhibited conditioned EEG responses during post-training sleep. However, selectively in trace-conditioned subjects, fear-conditioned cues also affected the valence of dreamed emotions. These findings suggest that hippocampus-dependent learning is accessible during non-REM sleep, and that hippocampus-mediated memory reactivation may be expressed, not only through neural activity in the sleeping brain, but also within concomitant subjective experience.

The likelihood of recalling dreaming and other cognition from sleep varies as a function of both the rapid eye movement (REM)/non-REM (NREM) cycles and the changes in cortical activation across the night. These processes determine cortical and subcortical patterns of regional activation in REM and NREM sleep, as well as the inhibition of afferent and efferent pathways within the brain. In sleep, activated regions of the cortex produce cognitive output similar to that generated during waking, but in the absence of sensory input, their output is dominated by a variety of biases, such as recent context and personal concerns.

It has long been known that dream recall, along with various other features of dreaming, changes as a function of time of night. Yet the processes which might account for these time-dependent variations remain obscure. Here we assess the contribution of homeostatic and circadian factors to the generation of NREM mentation across the diurnal cycle. Assuming that previously reported time-of-night mentation effects are primarily driven by a circadian activation cycle which approximates the core body temperature (CBT) rhythm, it was hypothesized that more content would be reported from daytime nap awakenings as compared to night awakenings. Afternoon Nap reports were compared to previously-collected nocturnal reports from Circadian Nadir and Late Morning time points. Contrary to our hypotheses, both amount of mentation reported and propensity to report any mentation at all were lower in Nap as compared to Late Morning reports. A purely circadian influence following the CBT cycle is inadequate to explain this pattern of mentation production.

The dual rhythm model of dreaming states that, under high sensory thresholds, heightened general cortical activation common to both REM/NREM and circadian-driven activation cycles sums to produce the main characteristics of dreaming. In addition, the unique pattern of regional brain activation characteristic of REM sleep amplifies the emotional intensity of the dream. Subjects were awakened from REM and NREM sleep once near the nadir of the core body temperature rhythm, where circadian-driven cortical activation was assumed to be low, and again in the late morning, where this activation was presumed to be high. As predicted, changes in the central characteristics of dream reports mirrored REM/NREM and circadian-driven fluctuations in general activation, while at the same time, the regional activation pattern unique to REM sleep amplified dream emotionality selectively in REM reports.

William Domhoff's The Scientific Study of Dreams breaks new ground in the field, not by proposing grandiose, premature answers to questions about the nature of dreams but rather by showing us with unprecedented clarity and scope where we have erred in the past and allowing us to start over again. In a careful, reasoned critique, Domhoff demonstrates the shortcomings of the dominant dream theories that haunted the last century. Based on a thorough examination of converging evidence from neurophysiological, cognitive, and content analysis approaches to the study of dreams, he then lays the groundwork for the construction of future theories of the dreaming process from a neurocognitive perspective. Although a new, complete theory is not presented here, its insightful and unbiased as sessment of where dream research stands at this time should make The Scientific Study of Dreams required reading for everyone interested an empirical approach to understanding dreaming. With the publication of Freud's Interpretation ofDreamsin 1900, popular thought about the nature of dreams became centered around the idea that our dreams are disguised wishes originating from unconscious influences in early childhood. Dreams allowed us to hallucinatorily gratify socially unacceptable wishes repressed in our unconscious, and the transformative process of dreamwork ensured that these themes were well enough disguised in our dream content to allow us to sleep through the night, without being startled awake. Until the early 1950s, this psychoanalytic conception dominated dream psychology and was explored largely using the same techniques of clinical observation that led Freud to the formula tion of his original theory. The modem era of empirical research on dreaming was kicked off in the 1950s with Aserinsky and Kleitman's (1953) groundbreaking discovery of rapid eye movement (REM) sleep. As a graduate student working on a summer project, Eugene Aserinsky serendipitously discovered that about every 90 min throughout the night, humans go through short periods during which their eyes dart back and forth underneath their eyelids, accompanied by dramatic increases in brain activity as measured by electroencephalography. Soon it was reported that sub jects were much more likely to report a dream when awakened from REM sleep

Although the emotional and motivational characteristics of dreaming have figured prominently in folk and psychoanalytic conceptions of dream production, emotions have rarely been systematically studied, and motivation, never. Because emotions during sleep lack the somatic components of waking emotions, and they change as the sleeper awakens, their properties are difficult to assess. Recent evidence of limbic system activation during REM sleep suggests a basis in brain architecture for the interaction of motivational and cognitive properties in dreaming. Motivational and emotional content in REM and NREM laboratory mentation reports from 25 participants were compared. Motivational and emotional content was significantly greater in REM than NREM sleep, even after controlling for the greater word count of REM reports.

communication or interaction between the processes or states of two or more neurocognitive subsystems, relative to the their interaction in the normal waking state. For the purposes of this discussion, we may separate these dissociations into three classes: 1. sensory transduction and afferent pathways, 2. brain processes including both explicit and implicit mental processes, and 3. efferent processes, including motor behavior, which includes rapid eye movements (REMs). For the purpose of this discussion we assume that dreaming is the cognitive component of a neurocognitive process that is particularly characteristic of Stage 1 REM sleep, and is intensified in the last hours of sleep as the brain is activated by both the 90 min REM-NREM cycle but also by the rising edge of the 24 hr. diurnal rhythm that supports the waking state. The dramatic dreams that most people remember tend to be the result of these two joint sources of activation (1). Sensory thresholds are generally elevated but also highly variable during REM sleep. When thresholds are high there is a marked dissociation between the sensory patterns that reach the sensory organs and the neurocognitive processes that take place in the association cortex. A variety of mechanisms account for this dissociation. Although eyelid closure accounts for part of the elevated threshold of visual perception, visual information is not transmitted even when the eyelids are taped open during dreaming sleep. Pompeiano (2) showed that the optic nerve transmitted little information, and Braun, Balkin, Wesensten, Carson, Varga, Baldwin, Selbie, Belenky, and Herscovitch (3), using H215O and PET to measure cerebral blood flow throughout the sleep-wake cycle, recently found that the activation of the visual projection region, the striate cortex, is actively diminished during REM sleep. In the auditory system, Pompeiano (2) also found that thresholds were increased by neural noise in the cochlea rather that high thresholds in the auditory nerve. At the efferent end of the system, the motor cortex delivers motor commands, but spinal inhibition fortunately prevents them from being executed (4,5). The respiratory and oculomotor systems are the only skeletal muscle systems to escape this inhibition. The former is clearly essential for survival, but the function of the latter is not known. Between the afferent and efferent systems lie the brain regions that produce the cognitive and affective characteristics of the dream. With the publication of the recent study by Braun et al. (3) we have reached the point where the measured patterns of brain activation account rather well for the salient characteristics of dreaming sleep. The paradoxical concurrence of the active mental experience of dreaming and the wake-like brain activity (electroencephalograms EEG) on the one hand, coupled with a bodily state that resembles a coma, baffled sleep investigators throughout the 20 years following Aserinsky and Kleitman’s discovery that dreaming sleep is associated

Recent work on functional brain architecture during dreaming provides invaluable clues for an understanding of dreaming, but identifying active brain regions during dreaming, together with their waking cognitive and cognitive functions, informs a model that accounts for only the grossest characteristics of dreaming. Improved dreaming models require cross discipline apprehension of what it is we want dreaming models to “explain.”[Hobson et al.; Neilsen; Revonsuo; Solms]

Using successive nonoverlapping 22.5-ms windows, 16-kHz sampling, no filtering, an automatic algorithm locates five successive windows of minimum spectral velocity, describes a subwindow equal to some multiple of the natural period of F0, and maps the subwindow onto the unit circle, the interval 0, 2π. Consequently, the Fourier analysis is performed on a window where the signal is exactly periodic. Because the spectrum contains no extraneous numerical sidebands it is precise, and consists only of natural harmonics in the acoustic signal. The first 32-integer multiples of F0 are sufficient to describe the spectrum. J. D. Miller’s [J. Acoust. Soc. Am. 85, 2114–2134 (1989)] log F01/3 shift increases recognition of 12 vowels (men, women, and children) in the Hillenbrand et al. [J. Acoust. Soc. Am. 97, 3099–3111 (1995)] data set from 52% to 75% using a Euclidean classifier (EC—with jackknife). Cosine series (12) were used to compare our Betancourt spectrum (EC: 76%) with the Hamming window (EC: 61%). Quadratic discriminant function analysis + log F0 (79%) adds only 3% to our best EC result. With this spectrum and a log F01/3 shift, most vowel information is clearly captured by a simple EC.

A set of automatic algorithms expresses the acoustic vowel signal as the ratio series, log(fj/F02/3), where fj are the first 32 integer multiples of F0, plus an additional 32 log(fj/F02/3) delta terms that represent vowel trajectories. F0 is measured on a window-by-window basis by an algorithm that eliminates all smearing due to conventional windowing algorithms. In order to reduce the dimensionality of this expression, the 64 terms are summarized by 11 + 11 terms from the cosine series. Using ten monosyllabic words spoken by 137 men, women, and children, vowel classification is within one percent of human accuracy. Because the model uses none of the circular definitions of formant measurement and makes only one assumption that is unique to speech, namely, the −log F01/3 offset, the log(fj/F02/3) contour is superior to a formant representation of voiced speech. Because a Euclidean classifier is as accurate as a quadratic discriminant function which uses more than ten times as many degrees of freedom, it is argued that the log(fj/F02/3) transform may be accomplished by a genetically acquired neural mapping of the acoustic signal that facilitates the learning of vowel categories by infants.