Visual hallucinations are defined as the visual perception of an object or event in the absence of an external stimulus. Presence of repeated visual hallucinations is an important clue to the diagnosis of DLB. Examination of visual hallucination is made by interview with caregivers, often utilizing a structured questionnaire such as the Neuropsychiatric Inventory (NPI) [37], as well as asking patient. However, when asking directly to the patient, the sensitivity would be low and the severity rating would be difficult because of the patient’s lack of awareness of hallucinations to be unreal or loss of memory for the experience of hallucinations. When interviewing caregivers, the different awareness and understanding of hallucinations of the caregivers would affect the assessment of the problem. Informants may often be absent. To overcome such problems, the pareidolia test has been developed as a tool directly evaluating visual hallucinations as like a performance on neuropsychological tests (Fig. 6.2) [38, 39].

Visual illusions are defined as the false visual perception of an object or event that is actually different. Although visual illusions are theoretically and conceptually different from visual hallucinations, visual illusions are very close experiences to visual hallucinations, are sometimes practically indistinguishable from visual hallucinations, and are also common in DLB. Among illusions, the experience of pareidolia is phenomenologically close to visual hallucinations. Pareidolias are illusions of meaningful objects such as faces and animals and are thought to arise from ambiguous forms embedded in visual scenes. For instance, patients with dementia with Lewy bodies may incorrectly see a person in a curtain or perceive blobs on the wall as faces. Uchiyama et al. [38] and Yokoi et al. [39] developed pareidolia stimuli with scenery pictures and noise figures to evoke and measure pareidolic illusions in patients with DLB and demonstrated that the number of pareidolic responses is correlated with the severity of visual hallucinations. The number of pareidolic responses on either version of the test was significantly greater in patients with DLB than in those with AD and healthy aged individuals, and the picture pareidolia test distinguished DLB from AD with a sensitivity of 100 % and a specificity of 88 %, and the noise pareidolia test with a sensitivity of 71 % and a specificity of 80 %. Cholinergic enhancement reduced the number of pareidolias on the latter test [39]. The pareidolia test can be used as a surrogate indicator of visual hallucinations in DLB and as a diagnostic tool for DLB [40].

Squire and Zola-Morgan [41] classified human memory into declarative memory and non-declarative memory; the former is memory that can be consciously recalled and described as an image or language, and the latter includes skills or procedural memory that cannot be consciously recalled as well as subconscious memory like priming and conditioning. Declarative memory is further divided into episodic memory and semantic memory. Episodic memory is the memory of specific events that individuals experienced, in which the episode is stored along with the temporal and spatial context, that is the situation when encountered. Semantic memory is equivalent to knowledge, i.e., organized memories for languages, concepts, and facts. Amnesia is a disorder of episodic memory, which includes anterograde amnesia where the ability to memorize new things is impaired and retrograde amnesia where ability to recall events that occurred before the development of the amnesia is impaired.

The human episodic memory based on the networks involving the medial temporal lobe including the hippocampus, thalamus, and basal forebrain and the lesions of any components in the networks lead to anterograde amnesia [42]. The medial limbic circuit (Papez circuit), i.e., hippocampus – fornix – mammillary body – mammillothalamic tract – anterior thalamic nucleus – anterior thalamic peduncle – cingulate gyrus (cingulate) – entorhinal cortex – hippocampus, is involved in encoding and consolidation; the ventrolateral limbic circuit, i.e., amygdala – ventral amygdalofugal tract – dorsomedial thalamic nucleus – anterior thalamic peduncle – basal forebrain/prefrontal cortex – uncinate fasciculus – amygdala, is believed to be involved in emotional memory (memory for emotional events). Other regions involving in episodic memory include the retrosplenial cortex, anterior temporal lobe, and prefrontal cortex.

It had been considered that degeneration of cholinergic neurons within the basal forebrain causes amnesia in AD [43, 44]. If that is the case, degeneration of cholinergic neurons of the basal forebrain might account for memory impairment in DLB, as DLB is characterized by a more profound degeneration of cholinergic neurons of the basal forebrain than in AD [45, 46]. However, the fact is that episodic memory is less impaired in DLB than in AD. Evidence is accumulating indicating that the medial temporal lobe including the hippocampus is responsible for memory impairment in AD [6, 10, 11, 47–50]. Atrophy of the medial temporal lobe is milder in DLB [51], which is consistent with milder impairment of episodic memory in DLB than AD.

Relatively preserved episodic memory is apparent not only on the everyday events but also on the memory tests. In general, recognition memory is disproportionately spared relative to both recall. The feature that episodic memory deficits are relatively mild suggests DLB rather than AD. Relatively milder memory impairment in DLB than in AD is evident on the memory items in the MMSE, DRS, and ADAS [6, 10, 47, 48], as well as on the various memory tests, for example, Wechsler Memory Scale-Revised (WMS-R) logical memory [11], California Verbal Learning Test (CVLT) [49], and Buschke Selective Reminding Test (BSRT) [21]. Visual memory is reportedly impaired more than verbal memory in DLB [21], which may be explained by the visuoperceptual deficits. Although impairment of episodic memory may not stand out in the early stage of the disease, it will eventually come out with the progression. In some patients with DLB, episodic memory is severely impaired, which is explained by the superimposed AD pathology on the DLB pathology [50].

Semantic memory and procedural memory are also impaired in AD. The anterior temporal lobe is crucial for semantic memory. Semantic memory is also relatively preserved in DLB [48]. Results of the category verbal fluency test and naming test that depends on semantic memory are comparably defective between DLB and AD [4, 21, 49, 50]. One report pointed out that phonetic and category verbal fluency in DLB was comparatively defective in DLB, while the former was relatively preserved than the latter in AD [48]. Procedural memory, the ability to learn skills that become automatic, involves the basal ganglia, cerebellum, and supplementary motor cortex. Procedural memory deficits are most commonly reported in patients with Parkinson disease among neurodegenerative diseases, independent of other cognitive dysfunction or dopaminergic medication [52, 53], while it has been reported that patients with mild AD can acquire motor, perceptual, and cognitive skills. [54] Therefore, procedural memory is likely to be defective in DLB and may be a distinctive feature from AD, although there were no studies addressing procedural memory in DLB.

Language is a typically localized and lateralized function, which is processed in several association areas including Wernicke’s area and Broca’s area usually located in the dominant hemisphere. As those areas are subject to be involved in dementing illness, impairment in language is a common finding among individuals with dementia and can be a presenting symptom, particularly in AD as well as primary progressive aphasia. Identification of language impairment is important in dementia, as it aids in the accurate diagnosis of a specific type of dementia, alters the prognosis, and changes the management.

In AD, with progression of the disease, naming and verbal fluency may be impaired. In some patients with AD, language disorder may disproportionally stand out, and moreover a type of primary progressive aphasia, logopenic progressive aphasia, may emerge [55]. As for DLB, language functions are largely intact, although there may be difficulties in finding words and understanding complex sentences. A few case reports described patients with DLB presenting with logopenic progressive aphasia, in whom AD pathology was superimposed [56–58]. Such a symptom is exceptional so far, and there are no reports of patients presenting logopenic progressive aphasia or any other type of primary progressive aphasia whose pathology is not compromised by AD pathology. It is unknown whether DLB presents with primary progressive aphasia.

Cognitive fluctuations with variation in attention and arousal, such as daytime sleepiness, staring spells, decreased awareness of surroundings, illogical thoughts, and incoherent behaviors, are a key feature of DLB. Although cognitive fluctuations are not specific for DLB, they are more prevalent in DLB; they occur in approximately 20 % of individuals with AD [59] and in 35–50 % of those with vascular dementia [60], while the prevalence increases to around 90 % in those with DLB [1]. Moreover, cognitive fluctuations in DLB have particular characteristics that are distinguishable from fluctuations occurring in AD. Bradshaw et al. [61] reported that fluctuating cognition in DLB had a spontaneous, periodic, transient quality, which appeared to reflect an interruption in the ongoing flow of awareness or attention that impacted on functional abilities, while fluctuations in AD frequently highlighted episodes of memory failure or a diminished capacity to cope with the cognitive demands of the immediate environment.

Cognitive fluctuations are not only of significant diagnostic importance but also of functional impact on patients in terms of significant independent effects on activities of daily living and increased care burden for caregivers [62]. However, their accurate identification and assessment present a major challenge. No operational criteria of cognitive fluctuations have been indicated in the diagnostic criteria, which is the most difficult to determine among the core symptoms of DLB [63]. There is no clear biomarker of cognitive fluctuations. Several instruments to measure cognitive fluctuations have been developed. Structured interview and questionnaire to caregivers, such as the Clinician Assessment of Fluctuation [64], One Day Fluctuation Assessment Scale [65], Mayo Fluctuations Composite Scale [66], Cognitive Fluctuation Inventory [67], and Dementia Cognitive Fluctuation Scale [68], are simple techniques and of clinical value to differentiate DLB from AD. The other technique to measure fluctuations is to utilize variance of simple reaction time and choice reaction time during attentional tasks. It measures the variation of the attention within a short period of time and can be a surrogate marker of fluctuating cognition [69], of which results correlated with those of the abovementioned structured interviews [69, 70]. These instruments are reportedly effective in differentiating DLB from AD [69, 70]. However, these have not been adequately tested as yet for reliability and validity, and the use of the reaction time tests in a clinical setting as a marker of cognitive fluctuations is limited because of the lack of available equipment and trained staffs [71].

The underlying mechanism of cognitive fluctuations is poorly understood. Which brain areas contribute to cognitive fluctuations is still unknown, and no obvious structural brain changes have been associated with cognitive fluctuations in DLB. Recent studies including functional neuroimaging studies suggested that cognitive fluctuations in DLB are associated with changes in the neural networks involving cholinergic and dopaminergic systems within the thalamus, which is the central to attentional function [72–75]. A cholinergic involvement in cognitive fluctuations in DLB is supported by results of a clinical trial of cholinesterase inhibitor, where donepezil significantly improved Cognitive Fluctuation Inventory score compared with placebo [16].