Belle Baum 2007
Belle Baum 2007
Belle Baum 2007
2007; 6: 184–192
ORIGINAL ARTICLE
Abstract
The cerebellum has long been considered to be mainly involved in motor function. In the last 20 years, evidence from
neuroimaging studies and from investigations of patients with cerebellar lesions has shown that the cerebellum plays a role
in a range of cognitive functions. While cerebellar contributions have been shown for learning and memory, the cerebellum
has also been linked to higher order cognitive control processes frequently referred to as executive functions. Although it is
widely accepted that the cerebellum contributes to cognitive processing, the nature of cerebellar involvement is not well
understood. The present paper focuses on the role of the cerebellum in executive processing, reviewing findings derived
from neuroimaging studies or from studies investigating deficits related to cerebellar dysfunction. As executive functions
cannot be considered as a unitary concept, special emphasis is put on cerebellar contributions to different aspects of
executive control such as working memory, multitasking or inhibition. Referring to models derived from motor control,
possible mechanisms of cerebellar involvement in executive processing are discussed. Finally, methodological problems in
assessing executive deficits in general and in assessing the cerebellar contribution to executive processing in particular are
addressed.
Correspondence: Dr Christian Bellebaum, Institute of Cognitive Neuroscience, Dept. of Neuropsychology, Ruhr-University of Bochum, Universitaetsstraße
150, D-44780 Bochum, Germany. E-mail: [email protected]
ISSN 1473-4222 print/ISSN 1473-4230 online # 2007 Taylor & Francis
DOI: 10.1080/14734220601169707
The cerebellum and executive control 185
There has been a long – and continuing – debate thalamus. In non-human primates, projections from
about the nature of the processes which constitute the frontal cortex to the pons have been shown to
executive control. Baddeley’s well-known model of originate mainly in the motor and premotor cortex,
working memory (14,15) comprises an attentional whereas a smaller proportion of fibres originates in
control system, the central executive, and two the prefrontal cortex (24–26). On their way to the
subsidiary systems for the storage of visuo-spatial cerebellum, the motor and the prefrontal loop
and verbal material, respectively. The central execu- remain anatomically segregated: Via the pons and
tive comes into play, when information has to be the cerebellar cortex, the motor cortex connects with
stored on-line for a short period of time and in the dorsal part and the prefrontal cortex connects
particular when this information has to be manipu- with the ventral part of the cerebellar dentate
lated (15). The concept of a unitary central nucleus (27). Consistent with cerebellar involvement
executive system has been questioned (16), partly in motor function, projections from motor and
because lesion and functional neuroimaging studies premotor cortex are much stronger in the primate
did not support the idea of a single brain region or compared to projections from prefrontal areas.
neuronal system as the substrate of executive Recent evidence has shown that the situation is
control. different in humans: Segmenting fibres contributing
More recent theoretical views conceptualise to the cerebral peduncle in terms of their cortical
executive control as a heterogenous concept made areas of origin with diffusion tensor imaging,
up of different subcomponents with potentially Ramnani et al. (28) could show that cerebellar
different neuronal substrates. Logan (17) defined projections from the prefrontal cortex (PFC) are
four executive subprocesses relevant for the follow- disproportionally stronger in humans compared to
ing behavioural circumstances: (i) Choice among non-human primates (28), suggesting that these
strategies to complete a task, (ii) instantiation of the connections have evolved quite rapidly in the course
chosen strategy, (iii) execution and maintenance of of evolution, similar to the PFC itself (21). In
the strategy, and (iv) inhibition of the strategy in accordance with the more prominent prefrontal loop
response to changing task demands. Along similar in humans, the ventral dentate is much larger in
lines, Norman and Shallice (18) developed the humans in comparison to the great apes (29).
concept of a supervisory attentional system (SAS)
and defined situations which require superordinate
Functional implications of fronto-cerebellar connectivity
executive control of behaviour. These are situations
which (i) involve planning and decision-making, As outlined above, to refer to executive functions as
(ii) involve trouble-shooting, (iii) require novel frontal lobe functions is inappropriate and an over-
sequences of action, (iv) are dangerous or technically simplification. Nevertheless, there is no doubt that
difficult or (v) require the overcoming of a strong the PFC is one of the key regions mediating
habitual response. In a more recent widely used executive processing (30–32). Phylogenetically, the
subdivision, Smith and Jonides distinguish five PFC is a relatively new structure which is more
executive processes operating on the contents of heavily interconnected with other regions than any
working memory: (i) Attention and inhibition, other structure. The PFC is supramodal (see 32),
(ii) task management, (iii) planning, (iv) monitoring which makes it an ideal candidate for the supervision
the contents of, and (v) coding representations in of lower level processing. Accumulating evidence
working memory (19). from lesion and functional neuroimaging studies
Taken together, executive control is not a unitary indicates that different regions within the PFC are
concept. Multiple subprocesses which are clearly related to different aspects of executive control, and
distinguishable on a behavioural level contribute to although the exact anatomy-function relationship
the overall concept, and each of the subcomponents remains to be determined, there is a consensus that
is likely to be related to a specific neural substrate. the lateral PFC plays a critical role. As there are
Although the approach of Smith and Jonides (1999) strong connections between the lateral PFC and the
has been very influential, in particular in functional cerebellum (see above), the cerebellum is likely to be
neuroimaging studies, there is as yet no consensus involved in the processing performed by this part of
about the number and the nature of executive the PFC.
subprocesses (11). A prominent deficit frequently observed following
lesions to the lateral PFC is reduced response
inhibition. Patients, e.g., show abnormally high
Fronto-cerebellar circuits
error rates in the interference condition of the
The cerebellum is thought to mediate cortical Stroop test (33), a task often used to assess the
information processing via closed cortico-cerebellar ability to inhibit prepotent response tendencies.
loops (11,20–23). Fibres originating from the cortex Comparable deficits of PFC lesion patients are
project via the pons to the cerebellum. From the observed across a range of different inhibition tasks
cerebellum, projections go back to the cortex via the (34,35), and a link between the lateral PFC and the
186 C. Bellebaum & I. Daum
ability to inhibit behaviour when necessary has also back to the cortex via the thalamus and later
been corroborated by functional neuroimaging compared with the actual sensory consequences.
studies (36). Another executive subprocess accord- Differences between the predictions and incoming
ing to the Smith and Jonides (1999) model which is information about the sensory consequences can be
closely linked to the lateral PFC is multitasking, i.e., used to update the internal model (54). The
the ability to efficiently coordinate different sensory advantages of an accurate internal model for the
inputs or behavioural outputs in parallel. PFC-lesion perceptual consequences of motor actions are quite
patients have problems in performing a digit span obvious. To influence ongoing behaviour, feedback
and a visual task at the same time, in performing two from the sensory systems and from proprioceptors
different decision tasks simultaneously or in com- arrives too late. An internal model can bridge this
pleting a variant of the Six Elements Test, which temporal gap. Also, an internal model allows a direct
requires the coordination of three different tasks comparison of the predicted and the desired state of
within a short time limit (13,37,38). Lateral PFC the motor system.
activations are also consistently observed in healthy Ramnani (2006) explicitly states that the very
subjects under multitasking conditions (39). same principles may be applied to more abstract
Multitasking is dependent upon a range of cognitive information processing and to non-motor function
capacities, i.e., learning and remembering rules (21). In general terms, a cerebellar internal forward
pertaining to the tasks, adequate planning and the model simulates information processing occurring in
ability to implement and to follow plans, with the all cortical areas to which the cerebellum is
dorsolateral PFC being most closely related to connected. Just as the cerebellum uses efference
planning (40). Other executive control subcompo- copies from the primary motor cortex to predict
nents related to the lateral PFC are set shifting, sensory consequences, efference copies from the
which has frequently been assessed using the premotor area coding for movement plans are used
Wisconsin Card Sorting Test (WCST) or variants to predict processing in the primary motor cortex.
of it (41–43), and working memory (44–49). With the information getting more and more
Taken together, a wide range of executive control abstract, loops originating in the PFC may be
subprocesses are associated with lateral PFC func- involved in predicting movement plans from more
tion, and the cerebellum is likely to be involved in general action plans (see 21). As the PFC is
the underlying processes via its reciprocal connec- connected with many different brain areas (see
tions with the lateral PFC. This may be particularly above), PFC-cerebellar circuits are likely to be
true for working memory, since the particular PFC involved in the generation of internal models for all
regions projecting to and receiving afferents from the kinds of processing, including cognitive domains
cerebellum are critically involved in working mem- such as memory or spatial perception. In accordance
ory (23). with this idea, interactions have been proposed
between the PFC, the cerebellum und temporal
and parietal areas in learning (55). The next
Mechanisms of cerebellar-cortical interactions
paragraph summarizes the current empirical evi-
In modelling the nature of cerebello-cortical inter- dence for a cerebellar contribution to executive
actions in information processing, many researchers control, followed by an integration of the findings in
focus on motor control (50). It has been suggested theoretical models outlined above.
that the cerebellum may play a similar role in
different behavioural contexts and that the same
Empirical evidence for cerebellar involvement in
processing mechanisms apply to motor and cognitive
executive functions
functions (21,51). This idea receives some support
from the cellular organisation of the cerebellum, In an attempt to systematically assess the cognitive
which is relatively similar over the whole cerebellar impairments following cerebellar lesions,
cortex (52). The processing performed by a parti- Schmahmann and Sherman (1998) examined 20
cular region of the cerebellum is thus mainly defined patients with cerebellar damage of different aetiolo-
by its connections with other parts of the brain. Ito gies. The authors reported executive function
(51) suggested that the cerebellum consists of deficits related to planning, set shifting, abstract
different microcomplexes, whose functions are reasoning, working memory and verbal fluency. The
defined by the unique connections to different areas authors referred to the complex deficit pattern,
of the cerebral cortex. In the motor domain, current which also included other cognitive impairments
models of cerebellar function propose that the cere- and personality changes, as ‘‘the cerebellar cognitive
bellum receives efference copies of neuronal motor affective syndrome’’ (56,57). Patients with involve-
commands from the cerebral cortex (21). With the ment of the posterior cerebellar lobe, i.e., the
help of an ‘‘internal forward model’’ the cerebellum territory of the posterior inferior cerebellar artery
is then able to predict the perceptual consequences (PICA), were found to be most severely affected.
of the movement (53). This information is relayed Although other studies corroborated some of the
The cerebellum and executive control 187
findings in cerebellar lesion patients, the data of cognitive processes apart from set-shifting and is
Schmahmann and Sherman (56) have to be inter- thus considered an ‘‘impure’’ test of executive
preted with caution, as lesion locations as well as function (66).
lesion – test intervals varied considerably. Most of The findings for task management or multitasking
the patients were tested in the acute phase after the as an explicit subcomponent in the theoretical
cerebellar lesion (earliest one week post-lesion) and framework by Smith and Jonides (1999) appear to
their performance improved, when they were re- be more consistent. Deficits were observed for the
examined 1–9 months later. Schmahmann and simultaneous performance of a movement and an
Sherman (56) related the deficits they observed to auditory task (67) and for a divided attention task
disturbed cortico-cerebellar interactions in informa- involving the simultaneous processing of visual and
tion processing and argued – in accordance with auditory stimuli (8), although this finding was not
ideas described above (21,51) – that the cerebellum replicated in another study of focal lesion patients
is linked to nearly every brain region and that there- (11). Consistent with this pattern, a recent PET
fore many different cognitive and affective distur- study also reported cerebellar activation in a dual
bances can accompany cerebellar damage. On the task compared to a single task condition in healthy
basis of their data, however, it was not possible to subjects (68).
relate specific cerebellar regions to specific deficits. The most frequently observed impairments
Focusing on the theoretical framework of execu- related to executive control which accompany
tive function subcomponents by Smith and Jonides cerebellar dysfunction in human subjects are deficits
(1999), findings on a cerebellar contribution to the in working memory and verbal fluency. As noted
inhibition of prepotent response tendencies are above, Schmahmann (1998) explicitly specified
inconsistent. In a follow-up study, patients showed impairments in these two domains as typical features
impairments relative to controls in the interference of the deficit pattern of patients with cerebellar
condition of the Stroop test, suggesting significant damage. Numerous neuroimaging studies yielded
inhibition problems in the acute phase of cerebellar cerebellar activations associated with working mem-
damage (lesion-test intervalv3 months). At one year ory tasks, in particular for verbal tasks, with
follow-up, however, patients’ performance had prominent right cerebellar hemisphere activations
improved significantly and did not differ from the (69–71), where spatial working memory did not
performance of control subjects (58). Similarly, yield cerebellar activations (72). Data from cerebel-
patients with selective cerebellar lesions in the lar lesion patients do not follow such a clear pattern,
chronic stage after ischemic damage showed intact with earlier studies yielding evidence of spared
inhibition abilities on the Stroop and a Continuous working memory as assessed with digit spans
Performance Test (11). In another study with post- (64,65), while more recent studies suggest impair-
acute patients, deficits in the Stroop test were ments (8,73–76). In the latter studies, verbal work-
confined to reduced verbal speed and did not affect ing memory was assessed with short word lists, a
inhibition (8). Data from functional neuroimaging verbal n-back task or digit spans. Along similar lines,
present a similarly inconsistent picture: Performance temporary disruption of cerebellar function by
of healthy human subjects was accompanied by transcranial magnetic stimulation (TMS) also led
cerebellar activation in an antisaccade task (59), but to verbal working memory impairments as assessed
not in a Go/NoGo task (60). It should be noted that with a Sternberg task (77). It should be noted in this
schizophrenic patients with problems in inhibiting context, however, that the working memory pro-
the pro-saccade showed similar cerebellar activation blems observed after cerebellar dysfunction are very
as normal subjects (59); it is thus unclear whether mild. Ravizza et al. (74), e.g., reported that the
cerebellar activation reflects processes related to patients’ scores in a digit span task, despite being
inhibition. significantly lower than the scores of control
For set shifting, studies with cerebellar lesion subjects, were in the normal range when compared
patients do not provide clear evidence for cerebellar to test norms. A transient verbal working memory
involvement, either. Consistent with cerebellar deficit has been observed in a single patient with a
activations during the performance of the WCST right cerebellar-lesion (76). Cerebellar TMS during
in healthy human subjects (61,62), Ravizza and Ivry the encoding phase was found to affect RTs but not
(63) observed higher error rates in cerebellar lesion accuracy in a Sternberg Task (77). Although spatial
patients compared to controls, when attention had working memory was largely found to be intact in
to be switched between two modalities. On the other cerebellar patients (e.g., 8), there is some recent
hand, performance differences between cerebellar evidence of a mild impairment which is clearly less
patients and control subjects on variants of the pronounced than in verbal working memory (74).
WCST have not been consistently found In verbal fluency tasks, subjects have to produce
(8,56,64,65). It is noteworthy that the suitability of as many words of a specified category (semantic
the WCST as a measure of executive function is fluency) or beginning with a certain letter (phono-
increasingly questioned since it taps a range of logical fluency) as possible within a given time
188 C. Bellebaum & I. Daum
period (typically one minute). The focus is thus on memory as the only cognitive domain. In parallel to
speed of production, but a range of other cognitive the cerebellar internal model for movement control,
processes such as access to verbal storage systems, Desmond et al. (83) proposed that the two
monitoring of retrieved information etc. are also components which make up the phonological loop
involved. Lesion and functional neuroimaging stu- of working memory, the phonological store and the
dies provide evidence that the cerebellum is critically articulatory control process (14), are both mediated
involved in verbal fluency tasks (78–80). Consistent by the cerebellum. Projections from temporo-
with the verbal nature of the task, activations were parietal cortex to the cerebellum are associated with
observed in the right cerebellar hemisphere (con- input from the phonological store, whereas fronto-
nected to the left cerebral hemisphere). The cerebellar projections reflect input from the arti-
problems of cerebellar patients appear to be more culatory control process. The function of the
pronounced for phonological compared to semantic cerebellum is to compare the content of the
fluency (81,82). A spatial analogue is the so-called 5 phonological store with the output of subvocal
Point Test, where subjects are asked to draw as articulation (83). Evidence from a single patient
many different figures as possible when connecting 5 with a right-sided cerebellar lesion seems to support
given dots. Cerebellar patients were impaired on this the idea of cerebellar involvement in rehearsal
‘‘spatial fluency’’ task, suggesting that the fluency processes (76). On the other hand, the findings of
deficit is not limited to verbal material (8). Ravizza et al. (74) argue against a cerebellar role in
Taken together, the evidence for a role of the rehearsal, as cerebellar patients showed normal
cerebellum in the different subcomponents of word-length and articulatory suppression effect.
executive functions is inconsistent, particularly when Given the reduced phonological similarity effect in
the lesion data are considered. The most consistent cerebellar patients, the cerebellum does seem to be
support for a cerebellar involvement stems from related to phonological encoding (73). However,
working memory and verbal fluency data. Few lesion Ravizza et al. (74) argue that phonological similarity
studies performed detailed analyses on the relation- effects do not necessarily reflect encoding but may
ship between the exact lesion location and the also be related to rehearsal difficulty (84). Together
pattern of executive deficits. When lesion side was with their finding that the observed impairments
considered, more severe deficits were often seen for varied depending on the exact lesion location, they
right-sided cerebellar lesions (e.g., 8), but significant conclude that the cerebellum might contribute to
laterality effects were not always found (e.g., 74). As both rehearsal and storage. The exact contribution
concerns the severity of the deficits, the impairments of different parts of the cerebellum to working
observed following cerebellar dysfunction appear to memory remains to be determined in the future.
be generally less pronounced than the problems seen Concerning the ability to efficiently coordinate
after PFC damage. two or more tasks (‘‘multitasking’’), cerebellar
processing is thought to underlie the increasing
automaticity of a recently learned movement (67).
The nature of cerebellar involvement in executive control
Performing more than one task at a time requires
The mechanisms of cerebellar involvement in that at least one of the tasks involved can be
‘‘higher’’ cognitive processing in general and in performed more or less automatically, so that the
executive functions in particular are far from being limited attentional resources can be directed to the
understood. Nevertheless, it is likely that the other, non-automatic task. While these hypotheses
processing performed by the cerebellum as concerns on the nature of cerebellar processing in working
executive functions somehow differs from that of the memory and multi-tasking still need to be corrobo-
PFC. Lesion studies are well-suited to elucidate rated by convincing empirical evidence, ideas about
subtle differences in the nature of the contribution of cerebellar involvement in other executive subcom-
the different regions involved in fronto-cerebellar ponents are less specific and need to be specified in
circuits. Differences in the severity of deficits more detail in future work.
observed in patients with PFC and cerebellar
damage provide some evidence for differential roles
Methodological issues in assessment of executive function
of both structures. Verbal working memory, for
example, is severely impaired in patients with frontal Several problems arise in the assessment of executive
lobe lesions but only very mildly in cerebellar lesion functions, which are likely to contribute to discre-
patients (74). pancies of empirical findings on the cerebellar
As outlined above, general principles of cerebellar contribution to executive processing. As outlined
involvement in motor control have been described in above, executive function is not a unitary concept.
some detail, and it has been suggested that the The definition of executive subcomponents differs
cerebellum contributes to motor and non-motor between different theoretical approaches (Shallice,
function in a similar way (21,51,57,83). However, Smith and Jonides), and there is as yet no consensus
theoretical models have so far addressed working in this regard. Nevertheless, there are standard
The cerebellum and executive control 189
neuropsychological tests addressing specific aspects above, an example from the motor context may serve
of executive functions. In many empirical studies of to illustrate the importance of the second issue. In
cerebellar patients available so far, only one or two the monkey, acute cerebellar lesions lead to sys-
tests were administered, without reference to theo- tematic changes in saccade amplitude. Saccades are
retical framework or specification of the particular either hypo- or hypermetric depending on the exact
subcomponent which is addressed. Many of the lesion location (88,89). After some time, monkeys
conventional ‘‘executive function’’ tasks are with vermis-lesions show complete recovery, if
‘‘impure’’ in the sense that they do not only assess average saccade amplitude is considered. They do,
the function in question but a range of additional however, still show dysmetria, i.e., the variability of
cognitive processes (85). The most frequently used saccade amplitudes is significantly higher as com-
task, the WCST, has been shown to be multifactorial pared to the pre-lesion period, and individual
(66). In addition, the conventional tests often lack saccades are markedly imprecise. Similar mechan-
‘‘ecological validity’’. Referring to earlier work (86), isms might come into play in cognitive domains like
Burgess et al. (85) distinguished between two executive function, and processing changes accom-
aspects of ecological validity: Representativeness panying cerebellar lesions might not be obvious at
and generalizability. Representativeness refers to first sight and require sophisticated assessment
the issue, whether neuropsychological tasks reflect procedures.
conditions or situations encountered outside the A more general problem in the assessment of
laboratory. Generalizability refers to the prediction cognitive functions in cerebellar lesion patients is the
of performance in everyday situations by perfor- potential ‘‘contamination’’ or effect of motor deficits
mance on a cognitive test. The fact that many on cognitive outcome variables. Many tasks that are
conventional executive function tests do not seem to conventionally used require motor responses, for
be sensitive to the ‘‘real life’’ problems of lesion example speeded verbal output in fluency tasks.
patients, implies that they do not fulfil these criteria Deficits in motor control need to be considered
to a sufficient degree. Furthermore, the type of by appropriate experimental or statistical control
processing required to solve a particular executive procedures.
function task was in most cases not related to a Despite these methodological concerns, the most
specific PFC region. The WCST again serves as an convincing approach to elucidate the contribution of
example. Although earlier evidence suggested that the cerebellum to executive function is the investiga-
problems tapped by the WCST only occurred in tion of patients with cerebellar lesions. An important
conjunction with dorsolateral PFC damage and not issue in this regard is the diversity of patients which
in patients with orbitofrontal lesions (87), many later have been studied so far, in particular with respect to
studies were not successful in replicating this pattern aetiology of cerebellar damage. In cerebellar atrophy
(see 85). patients, damage might extend to extra-cerebellar
Given the difficulties in assessing executive func- regions. In cases of more selective damage as a result
tions per se, elucidating the nature of processing of infarcts, haematomas, tumor resections or cere-
performed by the cerebellum poses an even more bellitis, lesion size and location differ substantially.
difficult problem. Even if a certain sophisticated Taken together, the patient studies have been very
behavioural executive function task reliably tapped useful in drawing attention to the cognitive opera-
the nature of PFC processing, it does not necessarily tions that the cerebellum might be involved in, but a
follow that the same task would also reveal the type specification of the precise nature of these operations
of cerebellum-specific processing contributing to in executive control awaits future theoretical outlines
overall executive performance. and empirical support. Useful research strategies
The proposal that cerebellar processing in a should include a focus on selective cerebellar lesions
cognitive context is analogous to the cerebellar and sophisticated cognitive tasks driven by theore-
contribution to motor control (see 28) has led to tical views on executive function in humans which
the idea of a general deficit pattern in cerebellar have been elaborated in recent years by cognitive
patients affecting motor and non-motor function in a psychology approaches (see 11). As different anato-
similar way. In accordance with the common motor mically defined cerebellar modules are thought to
symptom of dysmetria, Schmahmann and colleagues underlie information processing in different cogni-
termed the cognitive aspects of cerebellar dysfunc- tive and thus also executive domains, the region of
tion ‘‘dysmetria of thought’’ (57). This term implies interest within the cerebellum depends on the
that cerebellar damage does not necessarily lead to a executive subprocess which is being explored.
complete breakdown of performance in a specific
cognitive domain, but that there may be subtle
Acknowledgements
processing changes, which are difficult to reveal. In
assessing such a deficit, it is important to (a) apply a Our work referred to in this paper was supported by
good test and (b) measure the relevant variable. the Deutsche Forschungsgemeinschaft (DFG; DA
While the first issue has already been addressed 259/9-1).
190 C. Bellebaum & I. Daum
excisions, temporal lobe excisions or amygdalo-hippocam- 66. Greve KW, Brooks J, Crouch JA, Williams MC, Rice WJ.
pectomy in man. Brain. 1996;119(Pt 5):1597–615. Factorial structure of the Wisconsin Card Sorting Test. Br J
45. Owen AM, Sahakian BJ, Semple J, Polkey CE, Robbins TW. Clin Psychol. 1997;36(Pt 2):283–5.
Visuo-spatial short-term recognition memory and learning 67. Lang CE, Bastian AJ. Cerebellar damage impairs automati-
after temporal lobe excisions, frontal lobe excisions or city of a recently practiced movement. J Neurophysiol.
amygdalo-hippocampectomy in man. Neuropsychologia. 2002;87:1336–47.
1995;33:1–24. 68. Collette F, Olivier L, Van der LM, Laureys S, Delfiore G,
46. Barcelo F, Suwazono S, Knight RT. Prefrontal modulation of Luxen A, Salmon E. Involvement of both prefrontal and
visual processing in humans. Nat Neurosci. 2000;3:399–403. inferior parietal cortex in dual-task performance. Brain Res
47. Muller NG, Machado L, Knight RT. Contributions of Cogn Brain Res. 2005;24:237–51.
subregions of the prefrontal cortex to working memory: 69. Chen SH, Desmond JE. Cerebrocerebellar networks during
Evidence from brain lesions in humans. J Cogn Neurosci. articulatory rehearsal and verbal working memory tasks.
2002;14:673–86. Neuroimage. 2005;24:332–8.
48. Cohen JD, Perlstein WM, Braver TS, Nystrom LE, Noll DC, 70. Kirschen MP, Chen SH, Schraedley-Desmond P,
Jonides J, Smith EE. Temporal dynamics of brain activation Desmond JE. Load- and practice-dependent increases in
during a working memory task. Nature. 1997;386:604–08. cerebro-cerebellar activation in verbal working memory: An
49. Courtney SM, Ungerleider LG, Keil K, Haxby JV. Object fMRI study. Neuroimage. 2005;24:462–72.
and spatial visual working memory activate separate neural 71. Schumacher EH, Lauber E, Awh E, Jonides J, Smith EE,
systems in human cortex. Cereb Cortex. 1996;6:39–49. Koeppe RA. PET evidence for an amodal verbal working
50. Apps R, Garwicz M. Anatomical and physiological founda- memory system. Neuroimage. 1996;3:79–88.
tions of cerebellar information processing. Nat Rev Neurosci. 72. Jonides J, Smith EE, Koeppe RA, Awh E, Minoshima S,
2005;6:297–311. Mintun MA. Spatial working memory in humans as revealed
51. Ito M. Movement and thought: identical control mechanisms by PET. Nature. 1993;363:623–5.
by the cerebellum. Trends Neurosci. 1993;16:448–50. 73. Justus T, Ravizza SM, Fiez JA, Ivry RB. Reduced phonolo-
52. Bloedel JR. Functional heterogeneity with structural homo- gical similarity effects in patients with damage to the
geneity: How does the cerebellum operate? Behav Brain Sci. cerebellum. Brain Lang. 2005;95:304–18.
1992;15:666–78. 74. Ravizza SM, McCormick CA, Schlerf JE, Justus T, Ivry RB,
53. Blakemore SJ, Sirigu A. Action prediction in the cerebellum Fiez JA. Cerebellar damage produces selective deficits in
verbal working memory. Brain. 2006;129:306–20.
and in the parietal lobe. Exp Brain Res. 2003;153:239–45.
75. Bürk K, Globas C, Bosch S, Klockgether T, Zuhlke C,
54. Wolpert DM, Ghahramani Z. Computational principles
Daum I, Dichgans J. Cognitive deficits in spinocerebellar
of movement neuroscience. Nat Neurosci. 2000;3Suppl:
ataxia type 1, 2, and 3. J Neurol. 2003;250:207–11.
1212–7.
76. Silveri MC, Di Betta AM, Filippini V, Leggio MG,
55. Ito M. Bases and implications of learning in the cerebellum –
Molinari M. Verbal short-term store-rehearsal system and
adaptive control and internal model mechanism. Prog Brain
the cerebellum. Evidence from a patient with a right
Res. 2005;148:95–109.
cerebellar lesion. Brain. 1998;121(Pt 11):2175–87.
56. Schmahmann JD, Sherman JC. The cerebellar cognitive
77. Desmond JE, Chen SH, Shieh PB. Cerebellar transcranial
affective syndrome. Brain. 1998;121(Pt 4):561–79.
magnetic stimulation impairs verbal working memory. Ann
57. Schmahmann JD. Disorders of the cerebellum: Ataxia,
Neurol. 2005;58:553–60.
dysmetria of thought, and the cerebellar cognitive affective
78. Hubrich-Ungureanu P, Kaemmerer N, Henn FA, Braus DF.
syndrome. J Neuropsychiatry Clin Neurosci. 2004;16:
Lateralized organization of the cerebellum in a silent verbal
367–78.
fluency task: A functional magnetic resonance imaging study
58. Neau JP, Arroyo-Anllo E, Bonnaud V, Ingrand P, Gil R. in healthy volunteers. Neurosci Lett. 2002;319:91–4.
Neuropsychological disturbances in cerebellar infarcts. Acta
79. Schlosser R, Hutchinson M, Joseffer S, Rusinek H,
Neurol Scand. 2000;102:363–70. Saarimaki A, Stevenson J, Dewey SL, Brodie JD.
59. Tu PC, Yang TH, Kuo WJ, Hsieh JC, Su TP. Neural Functional magnetic resonance imaging of human brain
correlates of antisaccade deficits in schizophrenia, an fMRI activity in a verbal fluency task. J Neurol Neurosurg
study. J Psychiatr Res. 2006;40:606–12. Psychiatry. 1998;64:492–8.
60. Watanabe J, Sugiura M, Sato K, Sato Y, Maeda Y, Matsue Y, 80. Appollonio IM, Grafman J, Schwartz V, Massaquoi S,
Fukuda H, Kawashima R. The human prefrontal and parietal Hallett M. Memory in patients with cerebellar degeneration.
association cortices are involved in NO-GO performances: An Neurology. 1993;43:1536–44.
event-related fMRI study. Neuroimage. 2002;17:1207–16. 81. Bürk K, Globas C, Bosch S, Graber S, Abele M, Brice A,
61. Berman KF, Ostrem JL, Randolph C, Gold J, Goldberg TE, Dichgans J, Daum I, Klockgether T. Cognitive deficits in
Coppola R, Carson RE, Herscovitch P, Weinberger DR. spinocerebellar ataxia 2. Brain. 1999;122(Pt 4):769–77.
Physiological activation of a cortical network during perfor- 82. Leggio MG, Silveri MC, Petrosini L, Molinari M.
mance of the Wisconsin Card Sorting Test: A positron Phonological grouping is specifically affected in cerebellar
emission tomography study. Neuropsychologia. 1995;33: patients: A verbal fluency study. J Neurol Neurosurg
1027–46. Psychiatry. 2000;69:102–06.
62. Lie CH, Specht K, Marshall JC, Fink GR. Using fMRI to 83. Desmond JE, Gabrieli JD, Wagner AD, Ginier BL,
decompose the neural processes underlying the Wisconsin Glover GH. Lobular patterns of cerebellar activation in
Card Sorting Test. Neuroimage. 2006;30:1038–49. verbal working-memory and finger-tapping tasks as revealed
63. Ravizza SM, Ivry RB. Comparison of the basal ganglia and by functional MRI. J Neurosci. 1997;17:9675–85.
cerebellum in shifting attention. J Cogn Neurosci. 2001;13: 84. Jones DM, Macken WJ, Nicholls AP. The phonological store
285–97. of working memory: Is it phonological and is it a store? J Exp
64. Daum I, Ackermann H, Schugens MM, Reimold C, Psychol Learn Mem Cogn. 2004;30:656–74.
Dichgans J, Birbaumer N. The cerebellum and cognitive 85. Burgess PW, Alderman N, Forbes C, Costello A, Coates LM,
functions in humans. Behav Neurosci. 1993;107:411–9. Dawson DR, Anderson ND, Gilbert SJ, Dumontheil I,
65. Fiez JA, Petersen SE, Cheney MK, Raichle ME. Impaired Channon S. The case for the development and use of
non-motor learning and error detection associated with ‘‘ecologically valid’’ measures of executive function in
cerebellar damage. A single case study. Brain. 1992;115(Pt experimental and clinical neuropsychology. J Int
1):155–78. Neuropsychol Soc. 2006;12:194–209.
192 C. Bellebaum & I. Daum
86. Kvavilashvili L, Ellis J. Ecological validity and the real-life/ 88. Barash S, Melikyan A, Sivakov A, Zhang M, Glickstein M,
laboratory controversy in memory research: A critical and Thier P. Saccadic dysmetria and adaptation after lesions of
historical review. History Philosophy Psychol. 2004;6:59–80. the cerebellar cortex. J Neurosci. 1999;19:10931–9.
87. Milner B. Effects of different brain regions on card sorting. 89. Büttner U, Fuhry L. Eye movements. Curr Opin Neurol.
Arch Neurol (Chicago). 2006;9:90–100. 1995;8:77–82.