While disorders of affect in PD-D are common and clinically significant, psychosis may be the most clinically significant NPS in PD, as it is associated with cognitive decline and the development of dementia, worsening of motor symptoms, caregiver burden, and institutionalization [4, 28, 29]. Psychosis is defined by the presence of either hallucinations (i.e. false sensory perception) or delusions (i.e. fixed false beliefs). While hallucinations are common in PD overall, affecting approximately 25-44% of patients, the prevalence in PD-D is markedly higher than in PD patients without dementia , ranging from 45 to 65% [4, 8, 10]. Hallucinations are generally visual, well formed, recurrent, and complex [10, 31-33]; however, auditory [31, 34], tactile , gustatory
, and olfactory [36, 37] hallucinations may occur as well. Delusions are not as common as hallucinations, and typically co-occur with hallucinations. The most frequent forms are the ‘phantom boarder’ phenomenon (the belief that a stranger is living in the patient’s home) and delusions of infidelity. The imposter phenomenon is rare but can occur. Prevalence rates for delusions are reported to be 17% in PD patients overall [6, 9] and 25-30% in PD-D patients [4, 8, 10].
PD-D patients may display agitation and other behavioural disturbances, often in the context of psychosis. While not as common as depression, anxiety, or apathy , a large treatment study found baseline agitation/aggression of some severity in 33% of PD-D patients, with a third of those experiencing clinically significant symptoms. Of patients who had some degree of agitation/ aggression, half of caregivers reported clinically significant distress . In addition, approximately 30% of PD-D patients were reported to have some irritability or emotional lability, over 20% had aberrant motor behaviour, and over 10% demonstrated disinhibited behaviours.
Disorders of sleep and wakefulness
Disorders of sleep and wakefulness may be the most common non-motor symptoms in PD. Up to 90% of patients report insomnia, hypersomnia, sleep fragmentation, sleep terrors, nightmares, nocturnal movements, or rapid eye movement (REM) sleep behaviour disorder (RBD) [6, 38-40]. The latter is characterized by loss of normal skeletal muscle atonia during REM sleep resulting in dream-enacting behaviour, with speaking, shouting, and prominent motor activity associated with vivid, frequently scary, dreams. Other sleep-wake cycle-related disorders in PD include restless legs syndrome (RLS) and periodic leg movements in sleep (PLMS). Patients with more advanced PD may have an increased frequency of obstructive or central sleep apnoea .
Excessive daytime sleepiness (EDS) or fatigue occurs in 15-50% of PD patients [42-44], and is more common in PD-D. Sudden-onset REM sleep (also known as daytime ‘sleep attacks’) may also occur. This has usually been reported in conjunction with dopamine agonist treatment [41, 45] and is expected to occur less in PD-D patients as they are generally not prescribed dopamine agonists.
A recent study suggests that not only is RBD associated with an increased risk of dementia, it may also predict the onset of PD-D. A group of 42 PD patients were given a polysomnogram at baseline and reassessed a mean of 4 years later. At baseline 27 patients had RBD, and at follow-up 48% of them had developed dementia while none of the remaining 13 patients without RBD at baseline had developed dementia. Loss of REM sleep atonia at baseline significantly predicted hallucinations and cognitive fluctuations, symptoms that commonly co-occur with PD-D, as well as the onset of dementia .
Compta et al.  compared sleep architecture in patients with PD and PD-D with controls and looked at levels of hypocretin-1 in the cerebrospinal fluid (CSF). Hypocretin cell loss in the hypothalamus was found in PD patients with advanced disease. Patients with PD and PD-D had higher scores on the Epworth Sleepiness Scale (ESS)  than controls, with PD-D patients more likely to score more than the cut-off score of 10. However, there were no differences in hypocretin-1 levels between the PD, PD-D, and control groups, and levels were unrelated to ESS scores or cognition. PD-D patients did exhibit slow dominant occipital frequency and/or loss of normal non-REM sleep architecture .