[1] 11w The relationship between alcoholic cerebellar degeneration and cognitive and emotional functioning

[1] 171w Although it is now widely acknowledged that the cerebellum contributes to the modulation of higher-order cognitive and emotional functions, this relationship has not been extensively explored in perhaps the largest group of individuals with cerebellar damage, chronic alcoholics. Localised damage to the cerebellum has been associated with a specific constellation of deficits and has been termed the 'cerebellar cognitive affective syndrome' (CCAS) [Schmahmann, J.D., Sherman, J.C., 1998. The cerebellar cognitive affective syndrome. Brain 121, 561-579]. The CCAS describes a profile of impairments, including deficits in executive functioning and visuospatial skills, language disruption and altered personality and affective behaviour. It is conceivable that the CCAS may also develop in a subgroup of alcoholics with alcoholic cerebellar degeneration and may in part account for a proportion of the cognitive and affective deficits commonly observed with the condition. While evidence has emerged supporting such a relationship, methodological limitations and the lack of theoretically driven investigation of the contribution of cerebellar dysfunction to cognitive and emotional functioning in chronic alcoholics, preclude definitive conclusions being drawn.

[1] 188w Severe chronic alcoholism has many deleterious effects on the individual and results in a range of neuropsychological and neurological abnormalities (Harper, 1998;Moselhy et al., 2001;Parsons, 1994Parsons, , 1998;;Torvik et al., 1982). Along with generalised cerebral atrophy, specific cerebellar degeneration occurs in a significant proportion of alcoholics and associated clinical signs, including dysmetria and ataxia, are frequently observed. Accumulating evidence now suggests that, aside from its established role in motor coordination, the cerebellum plays a modulatory role in higher-order cognitive and emotional functions. This modulation is thought to occur via the interactive cerebrocerebellar circuitry, which connects the cerebellum to associative and paralimbic cerebral areas (Heyder et al., 2004). A specific constellation of deficits has been observed in individuals with localised damage to the cerebellum which has been termed the 'cerebellar cognitive affective syndrome' (CCAS) (Schmahmann and Sherman, 1998). The CCAS describes a profile of impairments, including deficits in executive functioning and visuospatial skills, language disruption and disturbed personality and affective behaviour. This syndrome is thought to occur due to disruption of the circuitry connecting the cerebellum and other brain regions, particularly those thought to subserve cognitive and affective functions.

[2] 164w Despite the fact that alcoholic cerebellar degeneration is probably the most common form of cerebellar disease (Timmann-Braun and Diener, 2000), to date little research has investigated the contribution of cerebellar damage to the cognitive deficits observed in chronic alcoholics. While evidence has emerged supporting such a relationship (Hillbom et al., 1986), a number of methodological limitations, in association with the small number of previous studies that have examined the relationship, preclude definitive conclusions being drawn. One such limitation is the disparity between studies in the methods employed to measure and quantify cerebellar ataxia, which is frequently used as an index of cerebellar damage. Furthermore, important drinking variables, such as years of heavy drinking, as well as premorbid factors, such as level of education and premorbid intelligence, have not been appropriately matched between groups in some studies (e.g. Hillbom et al., 1986;Sullivan, 2003). Future investigations must more carefully consider these methodological issues and develop designs that allow for replication, so that convergent evidence may accrue.

[3] 102w A further limitation associated with a number of the previous studies in this area is their lack of theoretically guided neuropsychological test batteries and indeed, studies have either included a limited set of measures (Sullivan, 2003), or exhaustive batteries (Hillbom et al., 1986). It is conceivable that the CCAS may develop only in the subgroup of alcoholics with alcoholic cerebellar degeneration and may in part account for a proportion of the cognitive and affective deficits commonly observed in alcoholism more generally. Such a theoretically driven investigation of the contribution of cerebellar dysfunction to cognition in chronic alcoholics is yet to be conducted.

[4] 157w Alcohol abuse is a maladaptive pattern of alcohol use that leads to clinically significant impairment or distress and results in a failure to fulfil major role obligations, repeated use in physically dangerous situations, recurrent substance-related legal problems and/or continued use despite persistent interpersonal problems (DSM-IV-TR; American Psychiatric Association, 2000). Chronic alcohol abuse typically leads to alcohol dependence, which is manifested by tolerance and withdrawal (DSM-IV-TR; American Psychiatric Association, 2000). Chronic alcohol abuse is a major contributor to a number of adverse social, physical and mental health outcomes and the economic cost to society due to alcohol abuse and dependency is rated in the billions of dollars in Australia (Collins and Lapsley, 2002). Chronic alcohol abuse has both direct and indirect toxic effects on many organ systems and as a result is a major risk factor in the development of a number of chronic conditions, including liver and cardiovascular disease (Arria and Van Thiel, 1992;Smart and Mann, 1992).

[5] 77w It is now well established that the cerebellum is essential for the coordination of skilled movement, the control of posture and gait and the regulation of muscle tone (Holmes, 1939;Ropper and Brown, 2005). The cerebellum plays an important role in the integration of sensory perception and motor output and essentially permits comparison between the actual movement being performed and the desired one, while allowing for the coordinated and smooth correction of any disparity (Ropper and Brown, 2005).

[6] 123w The cerebellum (''little brain'') is located in the inferior posterior portion of the skull, superior to the brain stem. Due to its heavily convoluted structure, it contains almost 50 percent of all of the neurons in the brain, although it constitutes only around 10 percent of total brain volume, highlighting its complexity and suggesting the significance of its role (Margolis, 2001;Sultan et al., 2000). The cerebellum is traditionally divided into the anterior, posterior and flocculonodular lobes; however, the organ may also be divided parasagittally into medial (vermal), intermediate (paravermal) and lateral (hemispheres) zones (Schmahmann et al., 1999). The cerebellum is connected to the brain stem by and communicates via, three large pairs of nerve fibre bundles, the superior, middle and inferior cerebellar peduncles.

[7] 146w There are three distinct layers of the cerebellar cortex: these are from outermost to innermost; the molecular, Purkinje and granular layers. The cytoarchitecture of the cerebellum is highly uniform, especially the lattice-like arrangement of the granule cell axons and Purkinje cell dendrites, which form hundreds of thousands of synapses in the molecular layer. The Purkinje layer contains the cell bodies of the large Purkinje cells, which are the primary integrative neurons of the cerebellum and provide its sole output (Ropper and Brown, 2005;Voogd and Glickstein, 1998). Each Purkinje cell receives an excitatory input from one climbing fibre originating from the inferior olivary nucleus (O'Hearn and Molliver, 2001). The granular layer consists of small granule cells, larger Golgi cells and mossy fibres. Mossy fibres enter this layer from their origins primarily in the pontine nucleus and form excitatory synapses with many granule cells (Sultan et al., 2000).

[8] 104w Output from the cerebellar cortex converges on three deep pairs of cerebellar nuclei, the dentate nucleus, the interposed nucleus and the fastigial nucleus (Colin et al., 2002). These nuclei are responsible for sending excitatory output to the cerebral cortex, as well as to postural and vestibular centres in the brainstem. Projections from the dentate and interposed nuclei travel via the superior cerebellar peduncle and decussate at the level of the inferior colliculus before reaching the contralateral ventrolateral nucleus of the thalamus (Ropper and Brown, 2005). The ventral thalamus then projects to ipsilateral cortical areas, thus the cerebellum influences motor functions on the ipsilateral side.

[9] 262w A tripartite, topological organisation of functions is postulated within the cerebellum. The vestibulocerebellum is primarily involved in equilibrium. The flocculonodular lobe and posterior vermis receive proprioceptive input from the vestibular nuclei of the medulla and outputs via projections from the fastigial nucleus to the vestibular and reticular nuclei (O'Hearn and Molliver, 2001;Ropper and Brown, 2005). Damage to this area of the cerebellum often leads to difficulties with balance, posture and eye movements, such as nystagmus. The spinocerebellum is involved in the adjustment of ongoing limb movements, along with the control of muscle tone and posture. The anterior vermis and part of the posterior vermis receive somatosensory and kinaesthetic information from the muscles, skin and joints regarding proprioception via the spinocerebellar tracts (Ropper and Brown, 2005). This region projects to the interposed nucleus, which in turn influences ventromedial pathway tracts. Damage to the cerebellar vermis results in postural instability and ataxia of gait and body sway are common signs of damage to this region. The cerebrocerebellum involves a complex feedback circuit, which mediates movement, planning and coordination. The cerebellar hemispheres receive input from the contralateral motor and association cortices and output via the dentate nucleus back to motor and premotor cortices. Damage to the cerebellar hemispheres results in abnormal rate, rhythm and force of movements resulting in problems with ballistic movements, such as difficulties carrying out rapid alternating movements (dysdiadochokinesis), under-or overshooting a target (dysmetria), difficulties in coordinating multi-joint movements (resulting in decomposition of movement) and tremor when moving a limb towards a target (intention tremor) (O'Hearn and Molliver, 2001;Ropper and Brown, 2005).

[10] 94w Common causes of cerebellar dysfunction include stroke, disease, tumour, inherited disorders and exposure to a number of toxic agents. Damage to the cerebellum itself, or to any of its associated pathways can result in characteristic cerebellar disturbances. The specific signs of cerebellar disturbance vary as a function of the site of damage; however, a common difficulty is in the coordinated execution of voluntary movements (Turner, 2006). Disorders of equilibrium and gait, decreased muscle tone, dysarthric speech, imprecise, clumsy limb movements and oculomotor problems, are the hallmark characteristics of cerebellar dysfunction (Ropper and Brown, 2005).

[11] 142w The cerebellum's participation in the modulation of higher-order functions may be understood in terms of its anatomical connections with numerous areas of the brain. Indeed, an abundance of efferent and afferent connections provide the neuroanatomical foundation for cerebellar involvement in cognition. The largest projection to the cerebellum is via the pontine nuclei, which receive an enormous number of nerve fibres from the cerebral cortex. Input is received not only from primary sensory and motor areas, but also from higher-level integration areas. The parietal, temporal and frontal association areas integrate information resulting in highly complex and organised projections that travel to the basis pontis, which in turn projects exclusively to the contralateral cerebellar cortex through the middle cerebellar peduncle (Brodal, 1978;Schmahmann and Pandya, 1989, 1991, 1995). This corticopontocerebellar pathway (feedforward limb) is the primary means of communication between the cerebral cortex and cerebellum.

[12] 103w These cerebrocerebellar connections are in fact bidirectional and a cerebellar feedback loop also exists, which projects from the cerebellum via the dentate nucleus through the thalamus, not only to sensorimotor cortices, but also back to higher-level integration areas (Clower et al., 2001;Middleton andStrick, 1997, 2001;Schmahmann and Pandya, 1990;Stern, 1942). This pattern of reciprocal interconnectivity between the cerebellum and a number of cortical regions, suggests that the anatomic substrates necessary for modulation of higher functions do indeed exist (Heyder et al., 2004;Schmahmann, 1991). Cognitive deficits may emerge due to damage to the cerebellum itself, as well as to the disruption of the cerebrocerebellar connections.

[13] 139w Numerous connections between the cerebellum and subcortical areas have also been proposed (Aas and Brodal, 1988;Schmahmann, and Pandya, 1993;Vilensky and van Hoesen, 1981), which appear consistent with the cerebellum's role in the memory component of classically conditioned responses (Schmahmann, 1991). Further, anatomic and physiological data from animal studies suggest that the cerebellum contributes to the limbic circuitry and has direct bidirectional pathways with the hypothalamus (Anand et al., 1959;Brodal et al., 1991;Cavdar et al., 2001a, b;Haines and Dietrichs, 1984;Heath and Harper, 1974;Heath et al., 1978). More recent neuroimaging studies have revealed functional connectivity between the dentate nucleus and a number of subcortical structures including the thalamus, hypothalamus, insula and basal ganglia in humans (Allen et al., 2005). Such circuitry could provide the cerebellum with information concerning motivation and affect, implicating it in the regulation of autonomic and emotional behaviour.

[14] 111w Further evidence of the connectivity between the cerebellum and distant brain regions derives from observations of reversed cerebellar diaschisis, whereby hypoperfusion in the contralateral cerebral association areas, as well as the thalamus and basal ganglia, has been found in patients with cerebellar damage (Botez et al., 1991;Schmahmann and Sherman, 1998). Crossed cerebellar diaschisis has also been demonstrated to occur in patients with lesions of the cerebral association cortices, whereby cerebellar hypometabolism occurs in the contralateral hemisphere of the cerebellum (Di Piero et al., 1990;Pantano et al., 1986). Such diaschisis has been ascribed to functional disconnection of cerebrocerebellar pathways, resulting in metabolic depression in areas anatomically remote from the site of injury.

[15] 108w Along with this anatomical connectivity, the cellular array of the cerebellum makes it a suitable structure to perform complex parallel processing. Its columnar collection of large numbers of cells is such that narrow modules have access to a wide range of information on which a large number of computations may be simultaneously performed (Leiner et al., 1991). Each narrow parallel zone sends its projections to one rod-like structure in the thalamus, which in turn projects in a small unit to a specific column-like area in the cerebral cortex (Leiner et al., 1991). Such arrangement allows for a large amount of parallel discourse between the cerebellum and the cortex.

[1] 242w Cumulative evidence now suggests that the cerebellum plays a role in the modulation of higher-order cognitive and emotional functions via its connections with a number of areas of the brain. While chronic alcoholism is known to result in a range of cognitive impairments, a review of the literature revealed few studies that have systematically investigated the contribution of cerebellar degeneration to the aetiology of such impairments. Both the alcoholic cerebellar degeneration and cerebellar syndrome literature generally, have been plagued by issues concerning the quantification of cerebellar degeneration. The contribution of important individual, nutritional and drinking variables to the development of alcoholic cerebellar degeneration requires further investigation. Whether cognitive impairments exist in alcoholics with cerebellar degeneration above and beyond those observed in alcoholics without cerebellar degeneration is of interest. If chronic alcoholics with cerebellar degeneration do indeed display additional deficits, it is of interest to characterise these impairments and to delineate whether there is a unique contribution of such cerebellar disorder, which distinguishes it simply from disconnection from other components of the distributed circuit. Furthermore, the presence and severity of cognitive and affective indices of the CCAS has not been investigated in individuals with alcoholic cerebellar degeneration. Such systematic investigation in the alcoholic population is warranted and would contribute to the growing body of literature concerning the relationship between the cerebellum and cognition and emotion and would additionally offer new insights into the aetiology of the cognitive deficits associated with chronic alcohol abuse.

[1] 79w Another common effect of chronic alcohol abuse is alcohol-related neurologic disorder. While acute alcohol use and intoxication has well-documented deleterious effects in both alcoholic and non-alcoholic persons (Dufour and Fe Caces, 1993;Lehman et al., 1993), chronic alcohol abuse leads to a variety of long-term cerebral insults and consequent cognitive disturbances. Common alterations in brain structure and function, as well as impaired neuropsychological functioning, have been identified in alcoholics (Harper, 1998;Moselhy et al., 2001;Parsons, 1994Parsons, , 1998;;Torvik et al., 1982).

[2] 241w The Wernicke-Korsakoff syndrome (WKS) affects around 2-12 percent of alcoholics, with estimates varying across studies (Torvik, 1987;Victor et al., 1989). It is a disorder most often encountered in alcoholics that is due to a deficiency of thiamine, an essential vitamin for the normal metabolism and function of brain cells. The deficiency may result from decreased dietary intake, but also indirectly through impaired absorption, storage and utilisation of the vitamin (Langlais, 1995;Martin et al., 2003). The acute phase of the disorder (Wernicke's encephalopathy) is characterised by a triad of symptoms, including eye movement abnormalities (opthalmoplegia), confusion and poor muscular coordination (ataxia). However, not all of the classic symptoms are always demonstrable (Victor et al., 1989). The symptoms improve with the administration of thiamine (Meyer et al., 1985), although approximately 80 percent of individuals are left with lasting deficits (Korsakoff psychosis), characterised by anterograde amnesia (profound short-term memory deficits), retrograde amnesia (with a steep temporal gradient and relative sparing of remote memories) and in some cases widespread cognitive impairments, including visuoperceptual and problem-solving deficits (Albert et al., 1979;Butters and Granholm, 1987;Victor et al., 1989). Patients with WKS are classically passive, apathetic and display reduced affect and initiative (Dufour, 1993;Martin et al., 1986). Confabulation is not uncommon amongst WKS patients, although this becomes less prominent after the acute stage has passed (Jacobson and Lishman, 1990;Victor et al., 1989) and lack of insight into their mnestic and other deficits is typical (Knight and Longmore, 1994).

[3] 78w Characteristically, lesions are observed in the diencephalic regions, including the mammillary bodies and the dorsomedial nucleus of the thalamus (Charness and DeLaPaz, 1987;Squire et al., 1990;Torvik, 1987;Victor et al., 1989). However, many investigators argue that these areas are affected in both non-amnesic and amnesic alcoholics (Blansjaar et al., 1992;Davila et al., 1994;Shear et al., 1996) and that the neuronal loss in the anterior thalamic nuclei is the exclusive characteristic of WKS (Antunez et al., 1998;Harding et al., 2000).

[4] 149w Well-nourished chronic alcoholics who do not develop WKS have also been found to exhibit significant brain impairment. Indeed, chronic alcoholism results in widespread cortical atrophy, as evidenced by decreased brain weight, ventricular enlargement and visible widening of cortical sulci, particularly affecting the frontal and frontoparietotemporal areas (Harper, 1998;Moselhy et al., 2001;Torvik et al., 1982). The generalised brain shrinkage is largely accounted for by the loss of cerebral white matter (Harper et al., 2003;Kril et al., 1997). Some of this damage appears to be reversible with abstinence, particularly in the first few weeks after cessation of drinking (Carlen et al., 1986;Ron, 1987), however, recovery may occur for months up to years after the cessation of drinking (Cardenas et al., 2007;Grant et al., 1987;O'Neill et al., 2001). This finding suggests that some of the atrophic changes observed may be related to reversible effects upon myelin and axons (Knight and Longmore, 1994).

[5] 96w Various neuropsychological deficits are associated with chronic alcohol consumption in individuals who do not develop WKS (Grant et al., 1987) and indeed, mild cognitive impairment has been demonstrated in 50-70 percent of detoxified alcoholics (Martin et al., 1986). Specifically, impairments are commonly observed in the domains of attention, visual and verbal learning and memory, visuospatial abilities, psychomotor speed and higher-order executive functions (Parsons, 1994(Parsons, , 1998)). General intelligence is usually spared and visuospatial functions appear to be more vulnerable than are verbal ones (Bergman, 1987;Di Sclafani et al., 1995;Fein et al., 1990;Knight and Longmore, 1994;Tarter, 1975).

[6] 114w Chronic alcoholism appears to chiefly disrupt the higherorder functions which are usually ascribed to the frontal lobes and their associated connections with other cortical and subcortical regions. Alcoholics frequently demonstrate difficulties in higher-order functions such as abstraction and problem-solving and in the ability to plan, organise and regulate behaviour (Parsons, 1994(Parsons, , 1998;;Williams and Skinner, 1990). Further, impairments in temporal ordering, set shifting and working memory are also commonly found in detoxified alcoholics (Noel et al., 2001;Sullivan et al., 2002Sullivan et al., , 2000d)). Alcoholics have also been shown to exhibit impaired impulse control, lack of insight, difficulty adapting to change and perseverative responding (Evert and Oscar-Berman, 1995;Moselhy et al., 2001;Oscar-Berman and Hutner, 1993).

[7] 102w Some investigators have argued that there are in fact no subtypes of functional deficits in alcoholics (Tivis et al., 1995) and that inefficiency (Nixon and Bowlby, 1996;Nixon and Parsons, 1991) or limitation of cognitive resources (Smith and Oscar-Berman, 1992) underlie the common processing deficiencies observed. Certainly, impairments in visuospatial skills have typically been measured using tasks that involve higher-order processes such as planning and organisational skills (Beatty et al., 1996;Leber et al., 1981;Sullivan et al., 1997). Similarly, difficulties on learning tasks appear to stem from the use of inadequate encoding strategies, rather than specific memory problems per se (Butters and Granholm, 1987).

[8] 116w In addition to metabolic and neuropsychological changes in alcoholics, a further feature commonly observed in a subgroup of this population is ataxia. Ataxia (derived from the Greek word for ''disorderly'') is a term used to describe incoordination of movement (Bastian, 1997). It essentially results from an inability to coordinate the actions of different muscle groups and to therefore appropriately adjust movements and is characterised by the dyscoordination of movement, gait and visual tracking, loss of balance and motor speech difficulties (dysarthria) (Shill and Hallett, 2001). While ataxia can occur after disruption to a number of different motor and sensory regions, it most commonly results from damage to the cerebellum (Ropper and Brown, 2005;Shill and Hallett, 2001).

[9] 151w The agreement between the severity of cerebellar signs and the degree of cerebellar atrophy as revealed by neuroimaging techniques have proven variable, but the association has only been investigated infrequently. While some studies have revealed good correlation between cerebellar atrophy and the clinical signs of ataxia (Koller et al., 1981;Melgaard and Ahlgren 1986;Sullivan et al., 2000a), discordance has been observed (Carlen et al., 1986;Estrin 1987). Indeed, Diener et al. (1986) found that concordance between objective cerebellar signs and CT imaging of cerebellar atrophy was only observed in 28 percent of the patients with cerebellar disorders in their study and in patients with involvement of the cerebellar hemispheres and anterior lobe, clinical signs were actually more pronounced than the cerebellar atrophy revealed by CT. Conversely, Hillbom et al. (1986) established that in a number of subclinical cases, signs of cerebellar degeneration were present on CT without the existence of clear clinical signs.

[10] 292w An issue that has plagued this literature and contributed to the variability in agreement between clinical and more objective measures is the heterogeneity of methods used to quantify cerebellar ataxia. Some studies have used fully quantitative measures, such as measurements of postural instability using frequency and time (Diener et al., 1984;Sullivan et al., 2004;Trouillas et al., 1988;Wober et al., 1999), while others have adopted semi-quantitative approaches using rating scales (Deshmukh et al., 2002;Lou et al., 1995). To help counter this problem, the Ataxia Neuropharmacology Committee of the World Federation of Neurology has developed the International Cooperative Ataxia Rating Scale (ICARS) (Trouillas et al., 1997). The goal of this exercise was to quantify the classic symptoms of cerebellar ataxia, so that the scale could also be used in other patient groups with cerebellar syndromes. It is a 100point ''semi-quantitative'' (Trouillas et al., 1997) scale comprised of four compartmentalised unequally weighted subscores in the domains of postural and stance disorders, limb ataxia, dysarthria and oculomotor disorders, allowing for separate study of individual subscores. Inter-rater reliability of the ICARS component scores and total score has been found to be high to very high in a group of patients with genetically determined ataxias, with a Kendall's o of 0.994 for total score (Storey et al., 2004). Such reliability was achieved in the absence of any prior discussion of the scale or training in its use by the raters. Furthermore, a significant inverse correlation between total ICARS score and cerebellar volume (r ¼ À0.805, po0.0001) measured by three-dimensional MRI was found in a group of patients with chronic cerebellar disease (Richter et al., 2005). This inverse correlation held for upper and lower limb ataxia, ataxia of posture and gait and dysarthria, but not for the oculomotor subscore.

[11] 184w Perhaps the most common form of cerebellar dysfunction, alcoholic cerebellar degeneration is frequently observed in alcoholics after around 10 or more years of heavy drinking (Andersen, 2004;Timmann-Braun and Diener, 2000). Alcoholic cerebellar ataxia typically evolves gradually over weeks to months, but may also occur abruptly or progress over many years (Victor et al., 1958). The condition may improve after abstinence in some patients (Diener et al., 1984;Sullivan et al., 2000c). Alcoholic cerebellar degeneration is characterised by ataxia of stance and gait, whereby a wide-based gait with shortened steps is adopted to compensate for the loss of balance and unsteadiness experienced (Andersen, 2004;Timmann-Braun and Diener, 2000;Victor et al., 1958). Patients often stand with feet placed far apart and trunk shifted forward and postural imbalance is often present, especially with eye closure (Romberg's sign) (Mauritz et al., 1979;Wober et al., 1999). Dysdiadochokinesis, terminal dysmetria and decomposition of movements are observed in varying degrees (Decker et al., 1959). Speech can be slowed and slurred, however is not usually characterised by scanning (variability of pitch and loudness) and a horizontal nystagmus is sometimes present (Victor et al., 1958).

[12] 173w Estimates of the prevalence of cerebellar degeneration in chronic alcoholics based on clinical, autopsy and radiological data indicates that around one-quarter to onethird of these individuals have cerebellar degeneration; however, estimates between 10 and 68 percent have been reported based on the basis of clinical findings (Deshmukh et al., 2002;Hillbom et al., 1986;Melgaard and Ahlgren, 1986;Scholz et al., 1986). This variability probably reflects differences in subject variables such as age, years of chronic drinking, lifetime alcohol consumption and the method of quantifying the ataxia. Two large autopsy studies revealed cerebellar atrophy in 26.8-27.6 percent, respectively, of all examined alcoholics (Lindboe and Loberg, 1988;Torvik et al., 1982). Prevalence rates of cerebellar atrophy in chronic alcoholics as revealed by CT have also been variable and have been reported to be from as low as 28-29 percent (Diener et al., 1986;Haubek and Lee, 1979) to as high as 71 percent (Melgaard and Ahlgren, 1986). More recent MRI studies have revealed that 33 percent of chronic alcoholics without WKS have demonstrable cerebellar atrophy (Antunez et al., 1998).

[13] 187w The predominance of gait and stance ataxia in alcoholics with cerebellar degeneration suggests that such damage may be localised to the anterior portion of the vermis. Indeed, early literature concerning cerebellar degeneration in chronic alcoholics revealed a predominance of degenerative changes in the anterior superior vermis and hemispheres (Victor et al., 1959). Imaging studies have since revealed both grey and white matter deficits (Sullivan et al., 2000a) and hypometabolism (Gilman et al., 1990) in the anterior cerebellar vermis in patients with clinical evidence of cerebellar dysfunction. The reported incidence of cerebellar hemisphere atrophy is more variable and has been reported to be as high as 84 percent in alcoholics with clinical cerebellar signs (Hillbom et al., 1986). In chronic alcoholics with radiological evidence of cerebellar atrophy, vermal atrophy was evident in 81 percent, while 57 percent had demonstrable cerebellar hemisphere atrophy, the latter only present in individuals with vermal atrophy (Koller et al., 1981). Partial reversibility of this atrophy appears possible and indeed, significantly faster and greater recovery in cerebellar tissue volumes have been noted in abstainers compared to relapsers at 8-month follow-up (Cardenas et al., 2007).

[14] 134w The different cell types of the cerebellum also appear to be differentially susceptible to damage due to chronic alcoholism. The Purkinje cells are the most vulnerable and global reductions in Purkinje cell numbers (Phillips et al., 1987) and densities (Andersen, 2004;Ferrer et al., 1984;Victor et al., 1959) are commonly observed in the cerebella of alcoholics. Similarly, Purkinje cell loss has also been demonstrated in animals receiving ethanol treatment, both during treatment (Walker et al., 1980) and recovery periods (Phillips and Cragg, 1984), suggesting that withdrawal processes may contribute to the death of these neurons. Murine models of ethanol-induced cerebellar degeneration must be interpreted with caution however, as some researchers have observed a different pattern of cerebellar pathology in rats, with granule cells and interneurons being more vulnerable than Purkinje cells (Tavares et al., 1987).

[15] 215w The exact pathogenesis of alcoholic cerebellar degeneration remains elusive and a number of putative factors have emerged as likely contributors. While alcoholic neurotoxicity and individual factors almost certainly play a role, thiamine deficiency has received the most attention and is thought by many to be one of the primary contributors to alcoholic cerebellar degeneration (Baker et al., 1999;Langohr et al., 1981;Martin et al., 2003;Phillips et al., 1987). Indeed, prolonged malnutrition has been frequently observed to precede cerebellar signs (Thomson et al., 1987;Victor et al., 1958). Perhaps the most compelling evidence implicating thiamine deficiency in the aetiology of cerebellar degeneration is the finding that cerebellar degeneration is often associated with Wernicke's encephalopathy (Harper, 1983;Torvik et al., 1982). Indeed, cerebellar ataxia is one of the cardinal features of Wernicke's encephalopathy and cerebellar lesions were once considered pathognomic of WKS. Victor et al. (1989, p. 121) have gone as far as to say that Wernicke's disease and alcoholic cerebellar degeneration in fact ''represent the same disease process''. One study, for example, has reported that all patients with acute Wernicke's encephalopathy showed evidence of cerebellar shrinkage on MRI (Antunez et al., 1998). Similarly, Baker et al. (1999) in a large autopsy study found that all thiamine-deficient alcoholics exhibited a significant reduction in Purkinje cell density and molecular layer volume.

[16] 118w Despite these data, other evidence suggests a more modest relationship between thiamine deficiency and alcoholic cerebellar degeneration (Estrin, 1987;Scholz et al., 1986). In a large autopsy study, cerebellar atrophy was observed in 26.8 percent of all examined alcoholics and only 38.6 percent of alcoholics with WKS (Torvik et al., 1982). MRI assessment has also observed cerebellar hemisphere and vermal atrophy in chronic alcoholics without WKS (Sullivan et al., 2000a). Furthermore, other groups have found mammillary body ratings to be uncorrelated with cerebellar ratings in non-amnesic alcoholics (Davila et al., 1994;Shear et al., 1996). Additionally, clinical features of cerebellar dysfunction can be reversed by thiamine administration in some, but not all patients (Lehman et al., 1993;Victor et al., 1989).

[17] 99w The precise mechanism through which thiamine deficiency might mediate cerebellar degeneration remains elusive. Clearly, thiamine plays an important role in the metabolic processes of the cerebellum and indeed, reduced levels are observed in other degenerative cerebellar ataxias (Langohr et al., 1981;Poloni et al., 1992). However, administration of thiamine in such individuals is often of no therapeutic benefit (Poloni et al., 1992). Furthermore, thiamine deficiency in other populations (e.g. beriberi, anorexia nervosa) does not primarily result in cerebellar degeneration (Kril, 1996;Miwa et al., 2004), suggesting that thiamine deficiency may be only one of several factors that adversely affect the cerebellum.

[18] 221w Direct neurotoxicity of alcohol itself has also been implicated in the aetiology of cerebellar degeneration in alcoholics. The acute effects of alcohol on cerebellar functions are well known and experimental studies with rats have revealed that ethanol directly decreases the firing rate of Purkinje cells through GABAergic mechanisms (Palmer and Hoffer, 1990). Furthermore, climbing fibre input is reduced by the administration of ethanol in rats, due to inhibition of metabotropic glutamate receptors (Carta et al., 2006). The longer-term deleterious effects of alcohol and its metabolites may occur through a number of mechanisms. Acetaldehyde, a metabolite of alcohol, is highly cytotoxic and contributes to alcohol-related liver damage (Timmann-Braun and Diener, 2000) and also directly affects the integrity of the cells underlying brain function (Pratt et al., 1990). An increase in free radicals and a decrease in antioxidants may also place the cerebellum under oxidative stress (Nordmann et al., 1990). Impaired liver function also appears to contribute to brain impairment in alcoholics, possibly due to the liver's reduced capacity to eliminate toxins from the bloodstream, along with nutritional deficiencies that occur secondary to liver disease (Butterworth, 1995;Tarter and Edwards, 1986). However, the frequency of liver cirrhosis has been observed to be almost identical in alcoholics with and without cerebellar atrophy, suggesting that liver damage does not significantly contribute to cerebellar degeneration (Torvik, 1987).

[19] 184w Important drinking pattern variables, including age of onset, duration of heavy drinking, continuous versus periodic (binge) drinking, type of beverage, typical amount consumed per session, number of withdrawals and period of sobriety before examination, also influence the degree of brain injury (Bergman, 1987;Cardenas et al., 2007;Errico et al., 2002;Jacobson and Lishman, 1990;Nicolas et al., 1997;Oscar-Berman et al., 2004;Sullivan et al., 2000c). However, while these variables have been found to be correlated with measurements of brain injury, they are often found to have limited predictive power in regards to both neuropsychological test performance and the degree of brain atrophy as evidenced by neuroimaging (Bergman, 1987;Fein et al., 1990;Parsons and Stevens, 1986). Indeed, while lifetime alcohol consumption has been found to be a reasonable predictor of gait and balance deficits in some studies (Sullivan et al., 2000d;Wober et al., 1999), others have found alcohol consumption variables to be uncorrelated with cerebellar ataxia ratings (Estrin, 1987;Johnson-Greene et al., 1997). It therefore appears that, while alcohol consumption variables contribute to alcoholic cerebellar ataxia, it is not a direct dose-dependent phenomenon and clearly additional factors contribute to its evolution.

[20] 43w Such factors may include genetic variables, including individual differences in the vulnerability of the brain to alcohol and/or its metabolites (Parsons, 1998). Other individual susceptibility factors such as gender, environmental influences and sociodemographic variables may also contribute (Arria and Van Thiel, 1992;Dufour, 1993).

[21] 117w Age appears to be an important variable, with studies finding a greater level of atrophy in the brains of older alcoholics, independent of duration of alcohol consumption (Pfefferbaum et al., 1992). Indeed, significant declines in Purkinje cell densities and atrophy of the cerebellar vermis have been observed in older non-alcoholics (Torvik, 1987;Torvik et al., 1986). More recent studies have identified a modest negative association between age and total vermal area in non-alcoholics and a negative correlation between even modest daily alcohol consumption and vermal size in this group (Piguet et al., 2006). It appears that a number of factors, including those that antecede, coincide with and which are consequential to chronic alcohol abuse, contribute to cerebellar degeneration.

[22] 213w The notion that the cerebellum may contribute to the modulation of higher-order behaviour and cognitive functions has attracted considerable attention in recent years and several lines of evidence now suggest that the cerebellum's contribution is not limited solely to controlling the execution of movement (Schmahmann, 1991). Evidence has accumulated that implicates the cerebellum in the mediation of a number of conditioned reflexes (Logan and Grafton, 1995;Solomon et al., 1989;Topka et al., 1992;Yeo and Hesslow, 1998). Early theorists proposed that the cerebellum may also be involved in the learning of rapid and skilful movements (Marr, 1969). Evidence for this proposal has emerged from neuroimaging studies whereby the cerebellum is activated in the early stages of motor skill acquisition (Doyon et al., 1997;Jenkins et al., 1994;Seitz and Roland, 1992). Furthermore, it has been demonstrated that patients with cerebellar deficits exhibit impaired procedural and skilled movement learning (Molinari et al., 1997;Sanes et al., 1990). Activation of the cerebellum is also observed in the absence of any motor activity, such that when healthy subjects are required to perform ideational, or purely mental tasks, such as imagining performing a motor activity, the cerebellum is activated (Decety et al., 1990;Ryding et al., 1993). Furthermore, impairments in such tasks are evident in patients with cerebellar lesions (Kagerer et al., 1998).

[23] 208w The recent interest in the cerebellum's contribution to higher-order functions has been fuelled by a number of neuroimaging studies that have found that the cerebellum is active during performance of a number of cognitive tasks. Such studies have shown that the cerebellum is activated by a number of non-motor tasks such as attentional shifting (Allen et al., 1997), mental imagery (Parsons et al., 1995), cognitive planning (Kim et al., 1994), verbal fluency (Gourovitch et al., 2000), set-shifting (Berman et al., 1995), working memory (Desmond et al., 2003;Desmond et al., 1997;Hayter et al., 2007;Klingberg et al., 1995), language tasks (Booth et al., 2007;Fiez and Raichle, 1997;Klein et al., 1995;Petersen et al., 1989;Raichle et al., 1994) and visual (Andreasen et al., 1996a) and verbal memory tasks (Andreasen et al., 1995a(Andreasen et al., -c, 1996b;;Grasby et al., 1993). Activation generally occurs in the lateral cerebellum contralateral to the concurrently activated region of the cerebral cortex (Desmond and Fiez, 1998). A further observation of interest is the significant correlation between cerebellar volume and visuospatial, executive and memory functions, along with fine motor dexterity, in healthy subjects (Paradiso et al., 1997a;Szeszko et al., 2003). Such correlations were corrected for cerebral size and indicating that they are not simply an artefact of brain size.

[24] 233w Further evidence of the cerebellum's role in higher-order cognitive processing is derived from examination of such processes in individuals with cerebellar dysfunction. A number of developmental disorders that compromise specific cognitive functions, such as attention deficit hyperactivity disorder (Berquin et al., 1998;Mostofsky et al., 1998b), autism (Courchesne et al., 1998) and fragile-X syndrome (Mostofsky et al., 1998a), have been associated with decreased cerebellar volume. Further, decreased posterior vermis size has been observed to predict a significant proportion of the variance (10-23 percent) on measures of intelligence, visuospatial ability and executive functions in females with fragile-X syndrome (Mostofsky et al., 1998a). The likely contribution of concomitant cerebral dysfunction to such deficits however, makes firm conclusions difficult. Patients with circumscribed cerebellar lesions and degenerative cerebellar disorders have also been observed to exhibit impairments in higher-order functions. Studies have found that around 20-24 percent of patients with degenerative cerebellar disorders and isolated cerebellar infarcts have functionally significant cognitive impairments (Kalashnikova et al., 2005;Leroi et al., 2002). Evidence from studies of unilateral cerebellar damage suggest that neuropsychological functions usually ascribed to lesions of the cerebral hemisphere contralateral to the cerebellar damage are impaired in such patients (Botez-Marquard et al., 1994;Fiez et al., 1992;Hokkanen et al., 2006). While some of these impairments were measured using neuropsychological tests that require a motor response, these impairments were also evident on tests that did not include a component of motor execution.

[25] 168w Visuospatial deficits have been observed in patients with cerebellar degenerative disorder (Akshoomoff et al., 1992), olivopontocerebellar atrophy (Botez-Marquard and Botez, 1993;Kish et al., 1988), excised left cerebellar hemisphere tumours (Wallesch and Horn, 1990), unilateral left-sided cerebellar lesions (Hokkanen et al., 2006;Wallesch and Horn, 1990) and following left superior cerebellar artery infarction (Botez-Marquard et al., 1994). Furthermore, impairments in visuospatial organisation after phenytoin intoxication have been found to improve in association with improvements in ataxia (Botez et al., 1985). Other studies however, have found no impairments in selective visual attention, visual spatial attention and mental rotation of designs (Dimitrov et al., 1996) and visuospatial organisation and memory (Daum et al., 1993) in patients with cerebellar damage. Daum et al. (1993) posited that such contradictory literature could be partly explained by group differences in background variables across studies, along with the use of non-comparable tests. Dimitrov et al. (1996) concluded that the previous reports of visuospatial deficits in patients with cerebellar degeneration concern quite specific functions that require further specification.

[26] 125w Studies concerning the role of the cerebellum in nonmotor learning tasks are so far inconclusive. Impairments in immediate visual memory have been demonstrated in patients with cerebellar degenerative disorders (Bracke-Tolkmitt et al., 1989). Some studies have also found such patients to be impaired on paired associative learning (Bracke-Tolkmitt et al., 1989;Drepper et al., 1999) and complex narrative recall (Kish et al., 1988), while others have not observed such impairments when damage is restricted to the cerebellum (Daum et al., 1993). While impairments have been observed on paired associate learning and word list recall tasks, subsequent normal performance on cued recall and recognition tasks suggest that problems with more effortful memory tasks may be secondary to executive function difficulties (Abel et al., 2005;Appollonio et al., 1993).

[27] 199w Indeed, Appollonio et al. (1993) found that patients with selective cerebellar degeneration were only impaired on tasks requiring the use of executive functions, such as an initiation/perseveration test and tests of phonemic and semantic fluency. Other executive function deficits have been noted in patients with cerebellar degeneration, including increased planning times on the Tower of Hanoi Test (Grafman et al., 1992), increased perseverative errors on the Wisconsin Card Sorting Test (Abel et al., 2005) and impairments in concept formation (Kish et al., 1988). In a single case study of a man with a large cerebellar lesion, deficits in executive function, including difficulties in planning, abstract reasoning, set shifting and perseveration, were observed 4-months post injury and after the motor features had disappeared (Paulus et al., 2004). Impairments in attention (Kalashnikova et al., 2005) and working memory (Hokkanen et al., 2006) have also been observed in patients with exclusively cerebellar infarcts. Conversely, Daum et al. (1993) found no evidence of executive dysfunction in their sample of cerebellar patients, a finding they attributed to their careful matching of controls with respect to age, IQ and mood scores, as well as ensuring that the cerebellar damage did not encroach on brainstem structures.

[28] 65w While damage to the cerebellum is not strongly associated with language impairments, language problems such as agrammatism (Silveri et al., 1994) and impaired performance on verb-for-noun generation tasks (Fiez et al., 1992) have been identified in patients with right cerebellar infarction. Further, Kalashnikova et al. (2005) reported limitations in the construction of verbal utterances in 65 percent of their sample of patients with cerebellar infarcts.

[29] 22w Paraphasia, anomia and circumlocutory expression have also been observed in a patient with a large posterior cerebellar lesion (Paulus et al., 2004).

[30] 140w Each of the aforementioned studies do however feature methodological limitations, which limit the conclusions that can be drawn concerning the cerebellum's contribution to cognitive functions. Firstly, activation of functional imaging during cognitive tasks in healthy participants does not necessitate that performance will be demonstrably impaired after lesions to the corresponding area. Secondly, the rare occurrence of confined cerebellar damage means that reports have commonly been based on single cases or very small samples. Such patient reports have often been comprised of heterogenous populations, whereby cerebellar damage has occurred through differing aetiological mechanisms and in different anatomical locations. Further, patient studies of cerebellar damage may reflect encroachment of pathology on other brainstem structures (Daum et al., 1993). While no one study provides conclusive evidence for the cerebellum's contribution to higher-order cognitive functions, taken together however, the evidence is somewhat more compelling.

[31] 219w A number of putative hypotheses concerning the role of the cerebellum in cognitive and emotional modulation have been proposed. Some maintain that the cerebellum functions as an error-driven adaptive control mechanism, which detects divergence from an intended outcome and functions to prevent and correct this error, whether these be errors in thought or in movement (Courchesne and Allen, 1997;Ito, 1993). Others have argued that the cognitive difficulties observed in patients with cerebellar lesions can be understood in terms of difficulties in controlling and regulating the temporal patterns of movement and behaviour (Ivry, 1997). In the same manner that the cerebellum modulates the timing of movements, it also contributes to mental processes that require representations of temporal information. Paulin (1997) has suggested that the cerebellum constructs neural representations of moving systems and is involved in perceiving, controlling and imagining systems in motion. Bower (1997) holds that rather than being responsible for particular behavioural functions, the cerebellum is involved in monitoring and adjusting the acquisition of sensory data required by the rest of the nervous system and facilitates a range of functions in this manner. Thach (1996Thach ( , 1998) ) posits that the cerebellum is involved in context response linkage through trial and error learning and as such an experiential context is able to elicit a certain mental action plan.

[32] 235w On the basis of past clinical and research findings, Schmahmann (1991) proposed the 'dysmetria of thought' hypothesis. In much the same way as the cerebellum controls the ''rate, force, rhythm and accuracy of movements'', Schmahmann (1991Schmahmann ( , p. 1183) ) proposed that it may also act to adjust the ''speed, capacity, consistency and appropriateness of mental or cognitive processes''. The dysmetria of movement that results from cerebellar dysfunction, is characterised by difficulties in the ability to check and correct movement which may be equated to a similar inability to check and correct errors in thought and behaviour. Schmahmann postulated that the cerebellum plays an important role in the fluidity and accuracy of mental processes and behaviours through its integration of multiple streams of information from many functional domains, including motor, sensory, cognitive, affective and autonomic activity. Schmahmann (2004) has described the cerebellum as acting as a 'universal transform' that helps to modulate and maintain mental operations and behaviour around an optimal level, or homeostatic baseline. This suggestion is in agreement with Leiner and Leiner's (1997) theory that the cerebellum acts similarly to a multi-purpose computer, designed to smooth out the performance of mental operations. Cerebellar damage disrupts the universal cerebellar transform and leads to disturbance of the cerebrocerebellar circuitry. The impairment manifests as ataxia when the anterior cerebellum is primarily involved and as the CCAS when the lateral hemispheres and posterior vermis are damaged.

[33] 187w Despite extensive evidence from anatomical, physiological and functional neuroimaging studies of the cerebellum's involvement in higher-order processes, the literature has produced relatively few case descriptions that clearly and adequately demonstrate this relationship. In the attempt to address this issue Schmahmann and Sherman (1998) examined 20 patients with diseases confined to the cerebellum. They identified a constellation of deficits in these patients, including deficits in executive functioning, visuospatial cognition, language, personality and affective behaviours, all of which resulted in a general lowering of intellectual functioning. They named this collection of impairments the CCAS. Specifically, disturbances in executive functioning included deficient planning, setshifting, abstract reasoning, working memory and decreased verbal fluency. Impaired spatial cognition was characterised by visuospatial disorganisation and impaired visuospatial memory. Personality changes were characterised by flattening and blunting of affect, or disinhibited and inappropriate behaviour. Linguistic difficulties included dysprosodia, agrammatism and mild anomia. These core features of the CCAS occurred in the context of normal arousal and alertness, preserved remote episodic and semantic memory and with only minimal effect on new learning. Further, these deficits could not be explained by the associated impairment in motor control.

[34] 120w This pattern of abnormality is usually encountered in patients with dysfunction of the cerebral cortex particularly in disorders of the underlying subcortical areas. Schmahmann and Sherman (1998) proposed that this pattern of deficits can be theoretically understood in terms of the cerebellum's connections with these areas. It is postulated that the topological organisation of cerebellar motor functions may also apply to higher-order functions, such that the cerebellar hemispheres modulate higher-order cognitive functions, while the flocconodular lobe and vermis modulate emotion and affect (Joyal et al., 2004;Schmahmann, 1991). Aside from the plethora of studies already described supporting a role for the cerebellum in the modulation of higher-order cognitive processes, several other convergent lines of evidence support the notion of the CCAS.

[35] 171w The cerebellum's contribution to emotion and emotional disorders gained recognition in the 1970s. Electrical stimulation of the cerebellar cortex in patients with both psychiatric illness and intractable seizures was found to significantly ameliorate aggression, depression and psychotic symptoms, as well as to improve general mood and increase alertness (Cooper et al., 1978;Heath, 1977) and to alter physiology in the limbic lobe (Heath et al., 1978). More recently, transcranial magnetic stimulation of the cerebellum in healthy volunteers has been observed to modify recordings over the prefrontal cortex and increase mood and alertness (Schutter et al., 2003). Further studies involving healthy participants have implicated the cerebellum in facial emotion recognition (George et al., 1993). Induced transient sadness (George et al., 1995), disgust (Paradiso et al., 1997b) and happiness (Habel et al., 2005;Paradiso et al., 1997b) have each been observed to increase activity of the cerebellar vermis. Further, lactate-induced anxiety attacks have been observed to result in a significant increase in blood flow in the left anterior cerebellar vermal region (Reiman et al., 1989).

[36] 60w Increased incidence of emotional disorders and personality change have been observed in patients with degenerative cerebellar diseases (Leroi et al., 2002). Disorders in the control of affect have been observed following atrophy of the cerebellar vermis (Gutzmann and Ku¨hl, 1987) and episodic rage is sometimes observed as an early symptom of tumours that encroach on the midline cerebellum (Elliott, 1982).

[37] 124w Animal studies have also contributed to our understanding of the role of the cerebellum in the control of affective behaviour. Cerebellar lesions to the vermis and flocculonodular lobe in highly aggressive rhesus monkeys have been observed to significantly decrease these aggressive behaviours (Berman et al., 1978;Peters and Monjan, 1971). Further, the cerebellar vermis has been implicated in fear conditioning in animal models (Sacchetti et al., 2004;Supple et al., 1993). Reciprocal connections with the brainstem, limbic system and prefrontal cortex form a neuroanatomical foundation for the involvement of the cerebellum in emotional modulation (Konarski et al., 2005). Whether the cerebellum is an integral and influential part of the system, or simply a relay station in the circuitry requires further investigation (Schutter and van Honk, 2005).

[38] 267w The CCAS is also evident in children that have undergone cerebellar tumour resection (Levisohn et al., 2000) and interestingly, mutism and dysarthria are not uncommon sequelae following such tumour excision (Catsman-Berrevoets et al., 1999;Pollack, 2001;Pollack et al., 1995;van Dongen et al., 1994). Such language disturbances have been observed to occur alongside other neurobehavioural symptoms such as regressive personality changes, apathy, marked emotional lability, irritability, impulsivity, disinhibition and poor initiation of voluntary movement (Levisohn et al., 2000;Pollack, et al., 1995). Deficits in affect modulation have been found in children with extensive damage to the vermis, but not when the vermis was unaffected (Levisohn et al., 2000). Neuropsychological sequelae of cerebellar tumour resection in children include impairments in executive function, such as deficits in planning and sequencing, visuospatial function and verbal memory (Levisohn et al., 2000). Others have observed impairments in verbal intelligence and on the performance of complex language tasks after resection of right cerebellar tumours and problems on non-verbal tasks and with prosody after left cerebellar hemisphere lesions (Riva and Giorgi, 2000). These authors also noted irritability and autistic-like features in children with vermal involvement. Deficits in the ability to rapidly shift voluntary attention between sensory modalities have also been demonstrated in such patient groups (Akshoomoff and Courchesne, 1992). Most of these symptoms are transient; however ataxia and mild dysarthria persist suggesting that the observed mutism may be an extreme form of dysarthria (Van Calenbergh et al., 1995). Further, persistent decreased motor speed, as well as attentional and executive dysfunction, after treatment for childhood posterior fossa tumours have been detected in early adulthood (Ronning et al., 2005).

[39] 128w Studies of neuropsychiatric disorders add further credence to the CCAS theory. Structural and functional abnormalities of the cerebellum have been noted in patients with schizophrenia for many years. Kraepelin et al. (1919) described a cerebellar form of dementia praecox, characterised by disorders of equilibrium, adiadochokinesia and tremor. More recently, subtle indications of motor impairment, such as dysmetria, have been observed in patients with schizophrenia, even in the absence of neuroleptic treatment (Andreasen et al., 1998;Gupta et al., 1995;Ho et al., 2004). Furthermore, neuroimaging studies of patients with schizophrenia (Heath et al., 1979;Ichimiya et al., 2001;Joyal et al., 2004;Loeber et al., 2001;Okugawa et al., 2003) and affective disorders (Beyer and Krishnan, 2002;Neil et al., 2005;Soares and Mann, 1997) have revealed volumetric reductions in the cerebellum, primarily in the vermis.

[40] 286w The contribution of cerebellar dysfunction to the cognitive impairments associated with neuropsychiatric disorders has also received much recent attention. A number of studies have examined this relationship and reported correlations between cerebellar measures and neuropsychological impairments in patients with schizophrenia (Flashman et al., 1996;Ho et al., 2004;Nopoulos et al., 1999;Szeszko, et al., 2003) and depression (Bench et al., 1992). However, other literature suggests that the relationship between schizophrenia and the cerebellum is not clear and that factors such as comorbid alcohol abuse and even gender may have important roles to play (Deshmukh et al., 2002;James et al., 2004;Sullivan et al., 2000b;Sullivan et al., 2004). Additionally, cerebellar pathology has been associated with an increased risk of developing affective and psychotic illnesses, such as depression and schizophrenia, as well as with organic personality change (Gray et al., 1998;Leroi et al., 2001Leroi et al., , 2002;;Turner and Schiavetto, 2004). Andreasen et al. (1996b) have suggested that a disruption of the neural circuits linking the cortex, thalamus and cerebellum may contribute to the psychopathology of schizophrenia. In the same vein as Schmahmann (1991), they hypothesise that such a circuit of interconnected nodes coordinates the fluid execution of mental activity. Disruption of this circuit and the resultant difficulties in processing and responding to information, is thought to be the fundamental cognitive deficit of schizophrenia and is able to account for many of its symptoms including hallucinations, delusions, disorganised speech and behaviour, alogia, anhedonia and attentional impairment (Andreasen et al., 1998). Indeed, recent attempts to understand schizophrenia have veered away somewhat from strict localisation of anatomical abnormalities and their relationship to specific symptoms and have focused more on the abnormalities in fundamental cognitive processes and distributed circuits (Braff, 1993;Friston and Frith, 1995).

[41] 175w To date, few studies have examined the contribution of cerebellar dysfunction to the cognitive impairments frequently observed in alcoholics and none have specifically investigated whether the CCAS exists in a subgroup of alcoholics. In the first study to compare neuropsychological performance in alcoholics with and without cerebellar degeneration, Hillbom et al. (1986) found that alcoholics with both clinical and radiological signs of cerebellar degeneration showed more neuropsychological deficits than alcoholics with only radiological signs or those with no signs at all. The clinically diagnosed group demonstrated inferior performance to the group showing no cerebellar signs on all of the tests administered, including measures of general intelligence (Synonyms, Reasoning, Block Design), visual and verbal memory and learning and the Halstead-Reitan battery. While patients with liver cirrhosis and WKS were excluded and age and level of education were included in the analysis, it must be noted that the average duration of alcohol abuse observed was significantly longer in the group with clinical signs and a previous history of head trauma and seizure problems were also more frequent.

[42] 410w Other groups have found no relationship between cerebellar damage and cognitive functioning in alcoholics. Davila et al. (1994) found that alcoholics performed similarly to controls on tests of verbal and non-verbal declarative memory, after accounting for differences in level of education and premorbid intelligence and furthermore, cerebellar ratings did not correlate with such measures. Johnson-Greene et al. (1997) compared chronic alcoholics with and without clinical evidence of cerebellar degeneration on tests of cognitive function and upper limb coordination. The groups were comparable on alcohol consumption variables, such as lifetime alcohol consumption and years of heavy drinking; however, some of the patients with cerebellar degeneration reported past nutritional deficiencies. Both groups exhibited impaired performance when compared to normative samples on a neuropsychological test of executive functioning, the Category Test, however did not differ from each other. The cerebellar degeneration group did display greater impairment in upper limb coordination than alcoholics without cerebellar degeneration. Sullivan (2003) proposed that cerebrocerebellar circuitry disruption may be one of the principal brain mechanisms underlying the cognitive and motor dysfunction observed in alcoholics. Abnormalities in volumes of brain structures underlying these circuits have been observed in alcoholics, including pontine and thalamic volume shrinkage (Harding et al., 2000;Kril et al., 1997;Lehman et al., 1993;Sullivan et al., 2003). Sullivan (2003) found decreased volumes in a group of 25 non-amnesic alcoholics in all structures subserving corticopontocerebellar and cerebellothalamocortical circuits. The patterns of correlations were consistent with a dissociation of thalamic and pontine circuitry. Pontine volumes correlated with volumes of the anterior superior vermis and cerebellar hemispheres, but not cortical volumes. Thalamic volumes correlated with volumes in the cerebellar hemispheres, inferior posterior vermian lobule and the parietal cortex. Inferior posterior vermian lobule volumes additionally correlated with parietal, prefrontal and frontal cortical volumes. Further, retrospective analysis of neuropsychological test performance revealed that vermian and thalamic volumes were predictive of problem solving skills and that cerebellar hemispheric white matter volumes were predictive of visuospatial ability. Interestingly, prefrontal and parietal cortical volumes were not predictive of such neuropsychological measures. It must be noted that the alcoholic group had fewer years of education and lower estimated premorbid intelligence than the age-matched control group. The neuropsychological deficits observed may well have been due to abnormalities in the individual structures that underlie the circuitry, or due to a disconnection or disruption of this circuitry. While lesions in any part of the circuitry may compromise function, impairments may be exacerbated when multiple subcomponents of the system are compromised.

[43] 301w The results of the above studies concerning the role of the cerebellum in higher-order processes in alcoholics are equivocal and further research is warranted. While Hillbom et al. (1986) and Sullivan (2003) found evidence of an association between alcoholic cerebellar degeneration and neuropsychological indices, including memory and learning, problem solving skills and visuospatial abilities, Davila et al. (1994) and Johnson-Greene et al. (1997) did not find any relationship between the degree of cerebellar damage and cognitive functioning in their alcoholic samples. The inconsistencies of these findings may stem in part from methodological issues inherent in the studies. Such issues include failure to match groups on important variables, such as years of education and premorbid intelligence (Sullivan, 2003), duration of alcohol abuse and history of neurological events (Hillbom et al., 1986) and history of nutritional deficiencies (Johnson-Greene et al., 1997). While some of the studies cited did use statistical means to account for individual differences, these were generally in relation to variables such as age, education and premorbid intelligence and not to alcohol consumption variables. Premorbid factors such as family history of alcoholism and prenatal exposure to alcohol were also not considered in the previous studies. Prenatal exposure to alcohol has been found to decrease children's ability to maintain postural balance (Roebuck et al., 1998). Furthermore, children from pedigrees with a high incidence of alcoholism present with higher levels of postural sway compared to children from low-risk families (Hill and Steinhauer, 1993). Such children have been found to show increased total cerebellar volume and lesser age-related grey matter pruning compared to control subjects (Hill et al., 2007). It is important to consider the possibility that premorbid differences in the integrity of the cerebellum due to prenatal exposure to alcohol and genetic pedigree may influence the degree of clinical indices of cerebellar degeneration.

[44] 196w Furthermore, the neuropsychological test batteries employed do not appear to have been guided by theories related to cerebellar system dysfunction and predicted associated cognitive deficits. The neuropsychological data used in the Sullivan (2003) paper was derived from a larger earlier study (Sullivan et al., 2000d), however only two neuropsychological tests representing two dissociable cognitive components were analysed in the later study, suggesting a less than rigorous foundation on which to base conclusions. Other studies have used all-inclusive, larger batteries, without proffering an explanation regarding the reasoning behind test selection (Hillbom et al., 1986;Johnson-Greene et al., 1997). While this approach can be useful in demonstrating the presence of deficits, test selection based on more targeted theoretically driven predictions is required to better conceptualise the nature of the cognitive impairments observed. Finally, as with the cerebellar degeneration literature generally, differences in the quantification of cerebellar degeneration in the alcoholic population limits the comparability of results across studies. The investigation of the contribution of alcoholic cerebellar degeneration to the cognitive deficits frequently observed in alcoholics is in its infancy and the methodological issues associated with the few previous studies may be used to guide the development of future research.

[45] 213w A further important consideration is the degree to which cerebellar damage is associated with generalised cerebral damage and it may be that the greater cognitive impairments in alcoholics with cerebellar degeneration may simply reflect more widespread damage. Hillbom et al. (1986) for example, found that all patients with radiological signs of cerebellar degeneration had at least some widening of cerebral cortical sulci. However, they also found that 83 percent of patients with no signs of cerebellar degeneration also displayed some supratentorial cortical atrophy, albeit to a lesser extent. Further, while the exact contribution of thiamine deficiency to alcoholic cerebellar degeneration is far from established, it is clear that such deficiency does play some role. It might therefore be postulated that cerebellar degeneration may occur alongside damage to other regions of the brain susceptible to thiamine deficiency, such as the diencephalic structures. Indeed, a correlation between damage to these two regions is well recognised in WKS patients (Antunez et al., 1998;Baker et al., 1999). However, such a relationship has not been clearly established in chronic alcoholics without frank amnesia, with damage to the two areas being found to be uncorrelated (Davila et al., 1994;Shear et al., 1996). It must be noted that there is a lack of evidence investigating this relationship in non-amnesic alcoholics.

[46] 226w Investigation of the relationship between the integrity of the cerebellum and emotional function has not so far occurred in the alcoholic population. Emotional changes are commonly noted in chronic alcoholics and it is not uncommon for the associated apathy and flat affect to be mistaken for depression in these individuals (Bedi and Halikas, 1985). However, while it has been reported that alcoholic Korsakoff patients show severe deficits in emotional functioning (Oscar-Berman et al., 1990), other researchers have not identified deficits in emotional responsiveness, with Korsakoff patients responding similarly to former heavy drinkers and light drinkers (Douglas and Wilkinson, 1993). Recently detoxified alcoholics have been found to exhibit impaired recognition of emotional facial expression as compared to both nonpatient controls and to patients with obsessive-compulsive disorder (Kornreich et al., 2001a) and such deficits have been found to correlate with self-reported interpersonal difficulties (Kornreich et al., 2002). Enhanced fear recognition and abnormal responses to the facial expressions of anger and disgust have also been observed in alcoholic inpatients (Townshend and Duka, 2003) and this inaccurate decoding of anger and disgust has been found to persist with abstinence of 2 months and greater (Kornreich et al., 2001b). Increased variability in subjective emotion and reduced physiological arousal changes in response to emotionally salient film excerpts has also been found in detoxified alcoholics as compared to controls (Kornreich et al., 1998).

[47] 185w It must be conceded that the observed emotional facial expression decoding problems may in fact have predated the chronic alcohol abuse in these populations, reflecting premorbid difficulties with emotional regulation and intelligence. A family history of alcoholism (an accepted predictor of future alcoholism) has been found to correlate with failure to show an enhanced startle response during presentation of unpleasant photographs in young adults (Miranda et al., 2002). Further, individuals with a probable history of prenatal alcohol exposure have been found to be significantly impaired on a task of affective prosodic comprehension (Monnot et al., 2001). While the alcoholic group also displayed affective prosodic comprehension impairments, the prenatal alcohol exposed group displayed significantly greater impairments, regardless of whether they had ever been diagnosed with alcohol abuse. Emotional and personality changes commonly observed in patients with specific cerebellar dysfunction, such as increased apathy, flattening and blunting of affect and disinhibition, are also frequently observed in chronic alcoholics (Parsons et al., 1987). It may therefore be postulated that at least some of the alterations in emotional functioning noted in chronic alcoholics may be attributable to cerebellar dysfunction.