The Role of the Cerebellum in Cognition and Behavior: A Selective Review

Lesions of the midline area of the cerebellum, the vermis, are associated with disorders of the trunk, whereas lesions of the lateral areas, the hemispheres, produce limb asynergia.¹ Cerebellar diseases can be generally localized by their clinical features: Lesions in the flocculonodular lobe are seen to cause disequilibrium with ataxic gait, a wide-based stance, and nystagmus; lesions of the anterior lobes are associated with an even more impaired gait and abnormal coordinated movements of the lower limbs; lesions of the lateral posterior lobes are associated with hypotonia, dysmetria, dysarthric speech, and dysdiadochokinesia.¹

From an anatomical standpoint, it should not be surprising that the cerebellum may play a role in nonmotor brain functioning. Although the cerebellum constitutes only 10% of the total brain weight, it contains more than half of all the neurons in the brain.¹ The cerebellum is connected to the cerebrum via three cerebellar peduncles. There are connections, largely via the thalamus, to many brain areas relevant to cognition and behavior, including the dorsolateral prefrontal cortex, the medial frontal cortex, the parietal and superior temporal areas, the anterior cingulate, and the posterior hypothalamus.^2,3 There are also noradrenergic, serotonergic, and dopaminergic inputs to the cerebellum from brainstem nuclei.² Given these connections, a role for the cerebellum in nonmotor functioning would seem likely. Gao et al.⁴ recently suggested that the lateral cerebellum is involved in the acquisition and discrimination of sensory information. Behavioral aspects of the cerebellum have not been directly examined until recently.

Schmahman and Sherman,¹⁵ using bedside cognitive testing as well as neuropsychological testing in a group of 20 patients with isolated cerebellar disease, described a syndrome that included impaired spatial cognition, dysprosody, and anomia, as well as executive dysfunction with difficulties in planning, set-shifting, abstraction, working memory, and verbal fluency. Abnormalities of the posterior cerebellum, especially if bilateral, were particularly associated with these cognitive difficulties. Although this study detailed both bedside cognitive abnormalities and neuropsychological testing in subjects with isolated cerebellar lesions, the patient group was heterogeneous, including patients with various diseases of the cerebellum. Additionally, neuropsychological testing was analyzed by using z-scores, with no control group for comparison. Moderate to severe executive dysfunction was similarly found by Storey et al.¹⁶ in an Australian pedigree of spinocerebellar ataxia. Although this study assessed executive functioning by use of various measures, there were only 5 subjects who completed all of the neuropsychological testing, and a control group was again lacking. Subjects with cerebellar disease have been often found to have “frontal-like” cognitive impairment, with much more variable findings in the areas of visuospatial dysfunction, language, and memory (see more detailed review by Daum and Ackermann¹⁷).

The cerebellum may also be relevant in the cognition of normal subjects without overt cerebellar disease. Cerebellar size has been found to be weakly correlated with memory retention and to show a trend for correlation with general IQ, even when covaried for cerebral volume in normal subjects.¹⁸ The relatively weak associations suggest that the role of the cerebellum in the cognition of normal subjects may well be mediated through the cortical areas with which it is intimately linked. In functional neuroimaging studies of normal subjects, the cerebellum has been seen to activate in tasks involving learning and word generation.¹⁹ These cerebellar effects do not occur in isolation and are rarely the areas of the most robust change, suggesting that the role of the cerebellum in cognitive changes in normal subjects is mediated by cortical areas.

Apart from its potential role in “coordinating” movement and cognition, the cerebellum may also be implicated in emotional and behavioral control. Schmahman and Sherman¹⁵ found that in their group of patients with isolated cerebellar disease, particularly those with midline and vermal pathology, personality changes of either flattening of affect or disinhibited and inappropriate behavior were common. The lack of standardized measures of these behavioral changes in subjects and the lack of a control group make this conclusion rather tentative. On the other hand, in a controlled study by Kish et al.,²⁰ patients with olivopontocerebellar atrophy had significantly higher depression scores than control subjects, and depression correlated weakly with cognitive testing. Mayberg et al.²¹ found that induction of transient sadness in healthy volunteers and patients with depression was associated with increased cerebral blood flow in the cerebellar vermis. However, this was but one of the brain areas found to have changes in cerebral blood flow with induction of sadness, and it is difficult to ascertain the role that the cerebellum plays independently of cortical and limbic changes.

An earlier study by Heath et al.²² showed that anterior cerebellar electrode stimulation improved some refractory cases of depression, psychosis, and behavioral problems in patients with diagnoses of schizophrenia, depression, epilepsy, and organic brain syndrome. With the availability of pharmacological treatments, now the mainstay of treatment for depression and schizophrenia, these observations may be seen as historically interesting but of limited practical value. However, the emerging use of transcranial magnetic stimulation,²³ other methods such as vagal stimulation, and stereotactic surgery for refractory cases in psychiatry may refocus attention on this previous observation.

There has been a growing interest in the role of the cerebellum in schizophrenia. An uncontrolled study showed that young male patients with schizophrenia who were on medications but not using alcohol had a preponderance of mild lower-extremity cerebellar signs²⁴ suggesting cerebellar involvement. Additionally, abnormal smooth-pursuit eye tracking has been found to be more common in schizophrenic patients (off neuroleptics) than in control subjects.²⁵ The abnormal eye movements may well be related to cerebellar pathology, although it is likely that alternative cortical systems including frontal eye fields were also involved.²⁶ These studies did not control for cortical involvement.

Some structural imaging studies have found cerebellar atrophy in schizophrenia,^27–29 but others have failed to replicate this.^30–32 One study in fact showed hyperplasia of the vermis.³³ Differences in both inclusion criteria and imaging methods may have accounted for these differences in the results. More precise MRI volumetric measures will be instrumental in resolving this debate.

Postmortem pathological studies in schizophrenia have shown smaller vermal area compared with subjects with no psychiatric illness or with other psychiatric illnesses;³⁴ smaller Purkinje cell size;³⁵ and decreased linear density and increased surface density of Purkinje cells compared with age-matched controls.³⁶ The influence of chronic treatment was not considered in these limited sample studies, nor have they been replicated. Additionally, although the control subjects when living had had no known psychiatric illnesses, they had not been thoroughly screened for the absence of psychiatric problems. Firm conclusions on structural changes of the cerebellum in schizophrenia therefore cannot yet be made.

A functional neuroimaging study by Volkow et al.³⁷ suggested that individuals with schizophrenia have lower cerebellar metabolism compared with control subjects. In this study, the subjects with schizophrenia were receiving neuroleptics and the control subjects were not; therefore it is unclear whether the cerebellar hypometabolism in the schizophrenic subjects was related to the illness or the medication. Additionally, the role of concomitant cortical changes was not explored. An intriguing new study by Crespo-Facorro et al.³⁸ of Andreasen's group³⁹ has suggested that subjects with schizophrenia have less blood flow in the cerebellum than control subjects during the performance of a novel memory task. This group has suggested the presence of a “cognitive dysmetria” in schizophrenia patients that relates to their cerebellar activity, analogous to the motor dysmetrias demonstrated in cerebellar patients. Their findings also suggest involvement of cortical-thalamic-cerebellar loops, since the cerebellum was but one area of altered blood flow, in addition to the frontal cortex, thalamus, and other areas. The role of metabolism or blood flow of the cerebellum in isolation in schizophrenia remains unclear. Validation of the paradigm in subjects with known cerebellar disease will be important for testing the specificity of these findings.

With respect to bipolar disorder, there has been some suggestion of cerebellar atrophy in patients with bipolar disorder or mania,^28,40 and another study showed a trend to this effect in patients over the age of 50.³² The role of alcohol abuse, however, may be a confounder. In one of the studies,²⁸ only the subjects with concomitant bipolar disorder and alcohol abuse had smaller cerebellar dimensions or vermis than control subjects. The other studies^32,40 did not control for alcohol abuse. Anticonvulsant medication use may be an additional confound.

Autism has been associated with hypoplasia of lobules VI and VII of the cerebellar vermis in a study by Courchesne et al.,⁴¹ although this finding has not been consistently replicated. (A recent review by Courchesne and others⁴² has demonstrated that in several MRI studies, patients with autism may have two types of cerebellar pathology—hypoplasia and hyperplasia—of the posterior vermis.) Kates et al.⁴³ studied a pair of monozygous twins, one of whom met criteria for strictly defined autism and the other of whom showed constrictions in social interaction and play but did not meet these criteria. Smaller cerebellar vermis lobules VI and VII were found in the affected twin compared with the nonaffected twin, further suggesting a role for the cerebellum in autistic disorder; however, there were differences in other brain regions as well, making this conclusion tentative. A recent study has shown smaller volumes of the posterior inferior lobe of the cerebellum in children with attention-deficit/hyperactivity disorder than in age-matched control subjects, even adjusting for brain volume and IQ.⁴⁴ Adults with Down's syndrome have also been found to have smaller cerebellar volumes than age-matched control subjects, also controlling for total intracranial volume and total brain volume. This difference did not appear to change over time in a small subset of patients followed serially.⁴⁵ These studies had the benefit of both a control group and covariate analysis controlling for brain volume. Specificity for symptoms in these disorders is not addressed in these studies, and dissimilarities in clinical presentations across syndromes likewise have not been addressed.

The cerebellum also appears have relevance to mechanisms in aging and dementia. With aging, a 10% to 40% decrease in Purkinje cell layer⁴⁶ and a reduction in the area of the dorsal vermis⁴⁷ have been reported, suggesting the possibility that any functions (motor and nonmotor) that are subserved by the cerebellum may be affected to some degree by the aging process. The role of the cell loss in mental or postural stability has not yet been studied. Alcoholic dementia is one of the classic dementias associated with cerebellar atrophy.⁴⁸ Although alcoholic dementia is commonly complicated by medical comorbidity, patients with this illness may have more ataxia and stereotypic behavior changes but less overt cortical dysfunction (e.g., less anomia, less deterioration in cognitive status) than do those with Alzheimer's disease (AD).⁴⁸ In contrast, Kish et al.²⁰ found that patients with olivopontocerebellar atrophy (OPCA) demonstrate multiple deficits in intellect, memory, attention, language, and visuospatial and executive functions compared with a control group. It is unclear whether these cognitive skills deteriorate over time in this population and to what extent these subjects had subtle cortical involvements implicating other sites of involvement in the absence of MRI correlation. Thus, although both alcoholic dementia and OPCA are associated with cerebellar abnormalities, it is uncertain how static these deficits are, and specificity remains uncertain because cortical and subcortical areas are also involved.

The cerebellum is not considered to be a primary focus of pathology in AD. However, diffuse amyloid plaques and increased microglia (but an absence of neurofibrillary tangles) can be found in the cerebellum, usually later in the AD process.⁴⁹ Purkinje cell density is decreased, especially in familial AD.⁵⁰ Ishii et al.⁵¹ found decreased cerebellar metabolism in severe AD, and this decrease was correlated with Mini-Mental State Examination (MMSE) scores. It is important to note, however, that this association may be an artifact of the temporal and parietal hypometabolism in these same patients, since this correlation was not corrected for cortical hypometabolism. In one autopsy study by Barclay and Brady,⁵² gross cerebellar atrophy had been found on CT scan in 2/8 (25%) of subjects with mixed dementia, but in none of 15 subjects with AD or 14 with multi-infarct dementia (diagnoses confirmed at autopsy); in view of these results, cerebellar atrophy on CT was tentatively suggested as a marker for mixed dementia. If replicated, this could be most helpful clinically.

The cerebellum may be implicated in the behavioral aspects of dementia as well. Gutzmann and Kuhl⁵³ found that affective lability and emotional incontinence in dementia are associated with cerebellar atrophy, third ventricular width, and interhemispheric fissure width, but not with other measures of cortical atrophy. However, it was unclear how affective lability and emotional incontinence were quantified, despite a clear attempt at attaining a homogeneous sample. Meguro et al.⁵⁴ found that wandering in vascular dementia was associated with sparing of the metabolic rate in the cerebellum as well as frontal, left parietal, temporal-parietal-occipital, and left occipital areas of the cortex. This finding only tentatively points to a role of the cerebellum and may be due to reciprocal functional connections between the cerebellar and cortical areas. In contrast to the finding of hypometabolism in the cerebellum in severe AD,⁵¹ Dolan et al.⁵⁵ found that patients with cognitive impairment in depression show higher cerebellar blood flow in the vermis and less blood flow in the left medial frontal cortex than depressed patients without cognitive impairment. This effect appears to be related specifically to cognitive dysfunction, since the investigators controlled for depression severity. If this finding is replicated, cerebellar activation may help distinguish between AD and the cognitive impairment of depression.