Using the analysis of sleep as a dissociated state,
I will (tentatively) suggest a way in which stimulant and sedative drugs might work to have their different therapeutic effects on different psychiatric symptoms or disorders.
Broadly speaking, stimulants may improve states involving emotional blunting/ anhedonia and also movement disorders (such as Parkinson's disease and catatonia) originating in the basal ganglia benefit; while sedative agents may benefit psychotic symptoms such as hallucinations and thought disorder, and also conditions excessive emotionality.
Sleep phase should be defined by what parts of the brain, what brain functions, are actually asleep - i.e. that are being rested and restored - and not by the brain regions or functions that remain active.
Therefore REM/ Dreaming sleep is a misnomer - because it describes the brain function that remains active, when the key aspect ought to be that during REM/ dreaming sleep, movement is paralysed (muscles are flaccid - except for eye movement) and (presumably) the basal ganglia - which fine-tune movement and emotions - are also being rested and restored.
In Deep Sleep, the musculature (and presumably basal ganglia) remain functional, but the key fact is that the brain functions related to awareness are being rested and restored (because during Deep Sleep we lose awareness, and are unaware of time passing, and have no long-term memory for what - if anything - has subjectively been happening).
Thus (to simplify - no doubt other functions are involved):
Deep sleep is about restoration of brain regions and functions related to awareness.
REM/ Dreaming sleep is about restoration of brain regions and functions related to emotions (and movements) - since emotions depend on the basal ganglia.
I suggest that stimulants and sedative have two, non-symmetrical actions on sleep phase architecture:
Stimulants stimulate REM/ Dreaming sleep - and thereby 'switch-off' and rest the basal ganglia with the effect of improving mood and emotional responsivity (and reducing basal ganglia induced movement disorders - such as Parkinson's disease and catatonia).
Sedatives suppress REM/ Dreaming sleep - and thereby switch-off and rest awareness. If (as I believe) hallucinations and thought disorder can validly be conceptualised as awake-dreaming - or, dreams intruding into the alert state to induce hallucinations, and breaking-up the stream of consciousness to induce thought disorder - then psychosis is a pathology of awareness.
The suppression of REM sleep by sedative could have therapeutic benefits if there is pathological awareness (for instance awareness of dreams intruding into the awake state), or emotions are too strong: that is if emotional strength is excessive.
There might also be an indirect effect from sedatives of re-balancing the sleep architecture towards Deep sleep and away from REM/ Dreaming sleep; and this may serve to reduce hallucinations and thought disorder.
However, the existing available sedatives only tend to enhance the shallower levels of deep sleep (Stages 1 and 2) and typically do not produce a natural and ideally restorative sleep (Stages 3 and 4) so that patients report sleeping more hours, but seldom wake-up fully refreshed.
An agent both stimulant and sedative would be Electroconvulsive therapy (ECT).
ECTs putative actions include a generalised Grand Mal cortical seizure that also (probably) involves the basal ganglia; then a 're-booting' phase after the seizure as the brain recovers from the depolarisation of a Grand Mal fit; an immediate post-ictal sleep; and the probability of better quality (more restorative) sleep on the night following the ECT treatment
How these effects might interact to improve sleep architecture is highly tentative, but ECT seems to be able to produce deeper Deep sleep Phase (eg Stages 3 and 4) than normal drug sedatives - at any rate, reports of sleep after successful ECT are that it is restorative and satisfying in a way that is superior to sedative drugs.
Furthermore, the benefits of ECT in demotivated and anhedonic Psychotic Depression, and also Parkinson's disease and catatonia (plus cerebrospinal fluid measurements) strongly imply that ECT enhances dopamine activity in both the meso-limbic system and the basal ganglia (especially the substantia nigra).
I would therefore suggest that ECT operates to enhance both Deep and Dreaming sleep - giving it an all-round therapeutic effect which goes beyond any of the usual drugs.