Dreaming during REM sleep: autobiographically meaningful or a simple reflection of a Hebbian-based memory consolidation process?

REM sleep is a brain state with desynchronized activity, often linked to dreaming. Dreams during REM sleep are typically bizarre, emotionally intense, and fragmented, yet feel story-like and personally meaningful. Elements of real-life events, especially trauma, tend to reappear, but the dream plot is distorted. This review presents a new testable model, the Trace-Spur-model, based on Hebbian neural network theory. It proposes that dream bizarreness results from changes in synaptic strength, allowing hyper-associative mental activity that may help consolidate or extinguish waking learning. Dream recall offers only a fragmentary glimpse into these processes, biased by autobiographical memory.

Study at a glance

Abstract

REM sleep is a state of desynchronized electrophysiological activity of the brain. It is usually accompanied by mental activity characterized by a succession of complex visual experiences commonly referred to as dreaming. Although REM sleep and dreaming are not implicitly conjoined, when they co-occur, they have a very distinct phenomenology, as, typically, the dream plot is bizarre and incohesive which is mirrored in heightened brain activation coupled with strongly attenuated coherence levels. At the same time, owing to increased limbic system activity, REM sleep dreams are highly emotional. Moreover, concrete emotions are often unrelated to dream events. Nevertheless, REM sleep dreams are often subjectively perceived as story-like and autobiographically meaningful. Indeed, elements of salient life events, attachment figures, and personally relevant emotions, especially trauma, seem to have a higher probability of re-appearing in dreams, albeit the dream plot itself remains highly distorted. This has prompted several theories on the interpretability of dreams, some authors leaning towards dreams reflecting waking mentation, others suggesting complete dissociation between waking and dreaming, both sides not fully accounting for empirical findings. In this review, we provide an overview of recent findings on the factors mediating REM sleep neurophysiology and dream content. As a first step towards integration of conflicting research results, we introduce a testable model (Trace-Spur-model) based on Hebbian theory of neural networks, proposing that dream bizarreness is a function of state-related modulations in synaptic strength allowing for hyper-associative mental activity, possibly enabling either a restructuring and integrative consolidation or extinction of learning experiences acquired in waking. In this model, dreams are viewed as phenomenological expressions of this neurophysiologic activity where dream recall allows a fragmentary witnessing of such processes, similar to peeking into an enduring and complex networking system. However, the content of the recollected dream is probably strongly deterred by autobiographical memory bias, favoring those images we can form some sort of association with.