Article Text


Cerebral correlates of psychotic symptoms in Alzheimer's disease
  1. Michael S Megaa,b,c,
  2. Linda Leea,b,
  3. Ivo D Dinova,b,
  4. Fred Mishkind,
  5. Arthur W Togaa,b,
  6. Jeffrey L Cummingsa,c
  1. aDepartment of Neurology UCLA School of Medicine, UCLA School of Medicine, Los Angeles, California, USA, bLaboratory of Neuroimaging, cAlzheimer's Disease ResearchCenter, dDivision of Nuclear Medicine, Harbor-UCLA Medical Cmt Torrance, California, USA
  1. Dr Michael S Mega, Laboratory of Neuro Imaging, Department of Neurology, UCLA School of Medicine, 710 Westwood Plaza, Los Angeles, CA 90095–1769, USA mega{at}


BACKGROUND Psychotic symptoms are produced by distributed neuronal dysfunction. Abnormalities of reality testing and false inference implicate frontal lobe abnormalities.

OBJECTIVES To identify the functional imaging profile of patients with Alzheimer's disease manifesting psychotic symptoms as measured by single photon emission computed tomography (SPECT).

METHODS Twenty patients with Alzheimer's disease who had SPECT and clinical evaluations were divided into two equal groups with similar mini mental status examination (MMSE), age, sex, and the range of behaviours documented by the neuropsychiatric inventory (NPI), except delusions and hallucinations. SPECT studies, registered to a probabilistic anatomical atlas, were normalised across the combined group mean intensity level, and subjected to a voxel by voxel subtraction of the non-psychotic minus psychotic groups. Subvolume thresholding (SVT) corrected random lobar noise to produce a three dimensional functional significance map.

RESULTS The significance map showed lower regional perfusion in the right and left dorsolateral frontal, left anterior cingulate, and left ventral striatal regions along with the left pulvinar and dorsolateral parietal cortex, in the psychotic versus non-psychotic group.

CONCLUSION Patients with Alzheimer's disease who manifest psychosis may have disproportionate dysfunction of frontal lobes and related subcortical and parietal structures.

Statistics from

Psychosis is common in Alzheimer's disease: delusions affect 10% to 73% of patients with Alzheimer's disease1-10 and hallucinations 3% to 49%.5-13 Delusions are most prevalent in women9 and during the middle phase of the illness.2 5 14 Psychotic patients with Alzheimer's disease exhibit an accelerated cognitive decline.15-17Visual hallucinations in Alzheimer's disease are more common than auditory hallucinations and, if accompanied by a fluctuating course or extrapyramidal signs, may suggest dementia with Lewy bodies.18 Although delusions increase from the mild to moderate stage of Alzheimer's disease, hallucinations decline in more severely impaired patients.14 The high prevalence of psychotic symptoms allows mapping of regional defects associated with delusions and hallucinations. Psychotic symptoms are correlated with metabolic and perfusion abnormalities in frontal,19-21temporal,22-24 and parietal cortex20 in Alzheimer's disease. Exploring the neural basis of psychotic symptoms in dementia requires control over prevalent concomitant behaviours, a methodological refinement not pursued in previous studies.

Psychotic symptoms do not occur in isolation. Agitation, anxiety, and irritability may be associated with increasing psychosis depending on the stage of dementia in Alzheimer's disease.14 Aarsland et al 25 reported that psychotic behaviour accounted for nearly 25% of the variance in measures of aggression in 75 patients with Alzheimer's disease. Controlling for accompanying behaviours is necessary in functional imaging studies that attempt to map the neuronal basis of symptoms such as psychosis in any disease population. The neuropsychiatric inventory (NPI) is a care giver based instrument26 that evaluates 10 troublesome behaviours in patients with dementia and provides the means for controlling for a wide range of behaviours in a functional imaging subtraction paradigm. We used the NPI to isolate psychotic symptoms, while controlling for other abnormal behaviours, in an Alzheimer's disease population undergoing technetium-99m d, l, hexamethylpropeleneamine oxime (99mTc-HMPAO) single photon emission computed tomography (SPECT) and clinical evaluation.



Starting from a pool of 280 patients with Alzheimer's disease presenting to the University of California, Los Angeles (UCLA) Alzheimer's Disease Centre, 20 outpatients were selected who met all clinical criteria described below and enabled equal group means across demographic and behavioural domains of the psychotic and non-psychotic groups. All patients met National Institute of Neurological and Communicative Disorders and Stroke/Alzheimer's Disease and Related Disorders Association (NINCDS/ADRDA) criteria for probable or possible Alzheimer's disease.27 In addition, all patients had acquired persistent decline involving at least three of the following domains: language, memory, visuospatial skills, cognition (calculation, abstraction, judgment, etc), and emotion or personality.28 Severity of cognitive deficit was measured in all patients using the mini mental state examination (MMSE).29


Family members living with the patient were interviewed with the NPI after procedures previously described26 in which screening questions for each behaviour were first posed. The care giver was asked if the behaviour represented a change from that exhibited by the patient before the onset of the dementia and if it was present during the past month. If a positive response was obtained then the behavioural domain was explored with scripted questions focusing on specific features of the behavioural disturbance. The care givers were then asked to rate the behaviours; scores from 1–4 were obtained for the frequency and 1–3 for the severity of each behaviour (a composite score for each domain was the product of the frequency and severity subscores; maximum=12). The 10 domains assessed using the NPI are delusions, hallucinations, agitation, depression, anxiety, euphoria, apathy, disinhibition, irritability, and abnormal motor output. The MMSE was administered at the same time as the NPI.


For all patients, an intravenous line was placed and 15 minutes were allowed to pass for patients to regain a quiet comfortable state before the intravenous administration of 30 mCi 99mTc labelled HMPAO (Ceretec; Amersham, Arlington Heights, IL, USA). Room lights were dimmed and the rooms kept quiet to minimise visual and auditory stimuli during the 15 minute brain uptake phase after injection. About 1 hour after injection, during which time washout of the tracer from the brain had occurred, SPECT images of the brain were obtained using a Picker 3000XP SPECT scanner (Picker International, Inc, Cleveland, OH, USA) with low energy ultrahigh resolution fan beam collimators. Images were reconstructed by filtered back projection using a low pass filter, eighth order, with a spatial frequency cut off of 0.23 to 0.25 cycles/pixel. Transverse, sagittal, and coronal planes with a 128×128 matrix were generated. Pixel sizes were nominally 3.56×3.56 mm. Resolution of the system was about 6 mm full width at half maximum (FWHM).


Spatial alignment of all 20 SPECT datasets was accomplished via 12 parameter affine registration.30 All datasets were first aligned and averaged to a random SPECT target to obtain an average “composite SPECT” which in turn was aligned to the International Consortium of Human Brain Mapping (ICBM) probabilistic atlas.31 To minimise resampling of data, the two registration fields above were concatenated and applied to each SPECT dataset. The relative perfusion scans of each patient then underwent linear intensity normalisation, on a voxel by voxel basis, to the global mean intensity value of all 20 patients, thus equalising the mean intensities across all datasets. This normalisation step did not alter the intersubject data variance or the mean intergroup differences.

Once all normalised datasets were in the common ICBM atlas space, a voxel by voxel subtraction was conducted between the psychotic and non-psychotic groups. Subvolume thresholding (SVT) was used to create a statistical map of these subtraction results according to methods previously detailed.32 Briefly, SVT utilises the probabilistic anatomical partitioning of the ICBM atlas (regions include the frontal, parietal, temporal, insular, and occipital cortex, along with the putamen, caudate, thalamus, and cerebellum) to model the different regions as separate stationary random fields thereby accommodating non-uniform global brain activity. This novel approach is particularly well suited for the assessment of functional imaging studies in Alzheimer's disease as parietal and temporal regions may have different means and variances, across subjects, than frontal or subcortical regions given the pathological distribution of the disease.33 34 Ignoring these potential differences by modelling the entire dataset as a stationary random field, done by many other functional assessments, will obliterate disease specific variability.

After the location of voxels within a region of interest (ROI) has been assigned a Z score value, a significance level must be determined for voxels above a Z score threshold. The SVT local search within globally significant regions derived from the between group subtraction is corrected for multiple voxelwise testing to control for type I errors in assessing significance. For each of the voxels selected by SVT a Bonferroni correction is conducted by dividing the significance level associated with the Z score by the number of voxels constituting a single search (this voxel number is equal to the size of the FWHM of the scanner—6 mm)


Table 1 shows the demographic and behavioural profile of the two groups. The demographic profile and MMSE mean score of the patients in this study were similar to the larger patient pool except for the overrepresentation of women in our two groups. Delusions (p<0.001) and hallucinations (p<0.01) were the only behaviours that were significantly different between the two groups. All psychotic patients had delusions and half had hallucinations. Trends for lower MMSE (p=0.37), greater agitation (p=0.54), and anxiety (p=0.28), with lower aberrant motor behaviour (p=0.44), were present in the psychotic group compared with the non-psychotic group. The figure and table 2 show the Talairach atlas35 location of the peak significance for regions with significantly lower perfusion in the 10 patients with Alzheimer's disease who had psychotic symptoms compared with the 10 non-psychotic patients with Alzheimer's disease as reported by care givers. The regions showing significantly lower perfusion in the psychotic group included the left and right prefrontal, left striatum, and left parietal cortex.

Statistical map of the regional, Bonferroni corrected, analysis of the non-psychotic (n=10) minus psychotic (n=10) patients with Alzheimer's disease. The map shows Talairach atlas location of the hypoperfused regions derived from normalised 99mTc-HMPAO SPECT, associated with psychotic symptoms.  

Table 1

Demographic and behavioural profile for the psychotic and non-psychotic groups as reflected by the composite (frequency×severity) scores on the 10 behavioural domains captured by the neuropsychiatric inventory (NPI)

Table 2

The Talairach atlas35 location of brain regions' peak, Bonferroni corrected, significance for the statistical map shown in the figure.


Functional imaging is the best tool for exploring the neuronal basis of neuropsychiatric disorders in life. Functional in vivo dissection of neural systems is only possible with metabolic or perfusion imaging. Psychosis is a complex behavioural disorder that does not manifest from a single brain defect. Disordered reality testing and abnormal inferential thinking are fundamental to psychosis. Such a disorder implicates executive and internal monitoring defects, and abnormal assessment of the emotional relevance of stimuli. We found significant hypoperfusion in the dorsolateral frontal cortex bilaterally, the left anterior cingulate, ventral striatum, pulvinar, and dorsolateral parietal cortex in psychotic patients with Alzheimer's disease. Both right motor (BA 4, 6, and 8), left prefrontal (BA 11, 8, and 4), and cingulate regions were hypoperfused supporting a defect in motor planning and cognitive executive function as well as the cingulate attentional system.

The left dorsolateral frontal lobe integrates language based executive function36-39 whereas the right seems to subserve internal monitoring of veridical choice.40 41 The anterior cingulate coordinates executive and self monitoring systems to simultaneously operate on modality specific sensory and association networks42-44 while coordinating the posterior parietal and dorsolateral frontal attentional network.43 Outflow from the anterior cingulate is directed to the ventral striatum,45 also termed the limbic striatum.46 This medial frontal subcortical circuit may coordinate the integration of emotionally relevant tone with executive processing; dysfunction could result in emotionally charged aberrant beliefs. Lesions of the left orbitofrontal cortex produce spontaneous confabulation.47 48 Dysfunction of these reciprocally connected frontal networks is here associated with the occurrence of psychosis in Alzheimer's disease.

No single brain location will be the source of psychotic symptoms in Alzheimer's disease but the medial and dorsolateral frontal cortical networks, in conjunction with the anterior cingulate limbic circuits, subserve many of the functions that seem to disintegrate with increasing psychotic symptoms. These regions have been implicated in the current study. This study has a small sample size selected from nearly 300 patients in an effort to produce similar demographic groups and control concomitant abnormal behaviours associated with psychotic systems. The drawback of such a sample purification is the loss of representation of a general Alzheimer's disease population as reflected by the low representation of men in this study. Yet the purification of the groups was needed to isolate and balance behaviours. Perfect balance was not achieved, however, given the trend for lower MMSE, higher agitation, and anxiety, with lower aberrant motor behaviour in the psychotic group. Future studies should determine if the neuronal basis of psychosis is similar across diseases and what baseline functional imaging patterns might predict treatment response.


Support for this work was provided by an NIA career development award (K08AG100784) to MSM; an NIA Alzheimer's Disease Research Center grant (P50 AG16570); an Alzheimer's Disease Research Center of California grant; the Sidell-Kagan Foundation, and the Human Brain Project (NIMH/NIDA: P20MH/DA 52176, NSF (BIR9322434), NCRR (RR05956).


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles