The current study applied a novel technique to measure the thickness of the cortical mantle in ALCs and healthy NC participants, without a priori defined volumetric ROIs. This method provided an objective and sensitive assessment of alcohol’s effects across the whole brain by calculating group differences first as function of cortical thickness across the entire mantle followed by statistically guided, data-driven region-by-region comparison, as opposed to a straight a priori region-by-region comparison. This analysis showed that ALCs have decreased whole-brain cortical thickness as compared to NC participants. The data-driven approach of the current method suggests that an in vivo measurement of the human cortex reveals it to be thinner overall in ALCs as compared to NCs; a finding that is highly consistent with the neuropathological literature demonstrating conclusively that alcoholic brains are lighter overall than nonalcoholic brains (e.g., Harding et al., 1996
; Harper et al., 1987
). However, while some pathological studies have attributed low brain weight to a reduction in white matter volume (e.g., De la Monte, 1988
; Harper andKril, 1985
), the current data support evidence that cortical gray matter is also reduced in volume and thus may contribute to low brain weight (e.g., Jernigan et al., 1991
; Mann et al., 2001
; Pfefferbaum et al., 1995
Mapping the thickness of the cortex across the brain revealed regionally specific areas that were thinner in ALCs compared with NC participants. Specifically, there were pronounced bilateral decreases in cortical thickness in inferior frontal, superior frontal, lateral frontal, middle temporal, precentral sulcus, central sulcus, and medial occipital cortical regions in ALCs. The regional distribution of differences suggests that the frontal lobes are especially vulnerable to thinning, followed by temporal and occipital regions. Thus, the current results support Harper and Kril’s (1989)
early pathological work and subsequent suggestion that alcohol may preferentially target large pyramidal cells in frontal cortex.
These findings are consistent with other volumetric techniques in humans (e.g., Chanraud et al., 2007
; Dao-Castellana et al., 1998
; Gazdzinski et al., 2005
; Makris et al., 2008
; Pfefferbaum and Sullivan, 2002
; Pfefferbaum et al., 2001
) showing selective neuronal loss occurring most consistently in frontal cortex. In particular, our results are highly consistent with a recent voxel-based morphometry study of gray matter (and white matter) in alcohol-dependent individuals by Mechtcheriakov and colleagues (2007)
. This study documented atrophy in the precentral gyrus, middle frontal gyrus, dorsolateral-prefrontal cortex, subcallosal, orbitofrontal, and cingulate cortices (as well as the subcortical structures of the dorsal hippocampus and anterior thalamus and the cerebellum). The overlap of regions most susceptible to alcoholrelated pathology in this volumetric study as compared to those identified using the cortical thickness whole-brain approach taken in this investigation is striking and provides strong support for this emerging consensus of vulnerable regions within the frontal lobes using different morphological techniques.
The regions identified as showing the greatest reduction in cortical thickness are consistent also with the well-documented neuropsychological effects of alcohol. Investigations of recently detoxified individuals with chronic alcoholism have consistently shown significant impairments in the cognitive domains of executive function, nonverbal memory, visuospatial function, and gait and balance (e.g., Oscar-Berman, 2000
; Oscar-Berman and Marinkovic, 2007; Parsons, 1993
). Deficits in these functions are suggestive of dysfunction of the cerebellar cortex and vermis, posterior parietal lobe, and most significantly, frontal lobes (specifically, orbitofrontal and prefrontal cortex) (Sullivan et al., 2000
) and are consistent with the ROI identified in this investigation.
Last, this is the first study to date to relate severity of alcohol abuse to reduction in cortical thickness. Length of alcohol abuse was related to a greater reduction in cortical thickness in this sample. Self-report of alcoholic behaviors as assessed with the SMAST was used to quantify alcohol abuse severity.
The SMAST is aimed at identifying alcoholic dependency through the endorsement of alcoholic behaviors. It is assumed that higher scores on the SMAST are associated with higher levels of alcohol consumption, dependency, tolerance, and maladaptive behaviors that interfere with social/occupational functioning. Thus, high endorsement of such alcoholic behaviors was associated with greater reduction in cortical thickness among recovered alcoholics. Severity of alcohol abuse was associated with greater reduction in thickness across the entire cortical mantle, and analyses with global thickness and ROI demonstrated that these relationships were independent of age. Correlations were highly significant for those regions found to evidence the greatest group differences. For example, the measures of thickness in the inferior frontal region as well as in the precentral sulcus region were most strongly related to severity of alcoholism. This latter finding in particular might have far-reaching consequences, as it is possible that reduction in frontal cortical tissue leads to maladaptive alcoholic behaviors or that the behaviors themselves are associated with greater alcohol consumption, which then exerts a neurotoxic effect on frontal cortical tissue. The causality of the relationship should be explored in future investigations.
Follow-up regression analyses with global thickness revealed a quadratic relationship between length of abstinence from alcohol and cortical thickness. Global cortical thickness was higher in this cross-sectional sample of ALCs with maintained sobriety for up to 12 years, after which thickness across the cortical mantle began to decline. This analysis supports the idea that changes to brain structure may occur within the initial 2 years of abstinence, but a longitudinal study would need to address this possibility directly. While certainly in need of further examination in a larger sample, this analysis supports other investigations demonstrating the brain’s ability to recover tissue integrity and function with maintained abstinence, particularly during the initial 2 years of sobriety. Why cortical thickness was decreased in individuals with 12 years or more of maintained sobriety is unclear, but likely is related to other mediating factors rather than to alcoholism or abstinence per se. It is important to note that there are not many individuals in the sample who have achieved sobriety for more than 12 years.
Although our approach is a very sensitive and accurate way of measuring neuropathology in vivo, there are other brain structures that we did not investigate but are nonetheless impacted by alcohol, most importantly, the cerebellum. The cerebellum is in fact the primary neurologic target of alcohol’s deleterious effects. Given the large body of literature confirming cerebellar shrinkage among chronic and ALCs, if we had been able to visualize cerebellar cortex, we would predict that we would have seen alcohol-related thinning in the cerebellum as well. We continue to develop our ability to assess volume, cortex, and white matter in the cerebellum and hope to address the association between cerebellar and cortical atrophy in future investigations.
These findings demonstrate widespread reduction in cortical thickness as a consequence of alcoholism in a sample of community dwelling former alcoholics. It should be stressed that the current findings were derived using a whole-brain analysis that yielded regions of greatest significance, rather than an a priori region-by-region comparison that can be susceptible to variations in the statistical sensitivity of each isolated region/comparison (Fischl and Dale, 2000
). This approach has advantages in detecting regionally specific cortical atrophy associated with subtle brain changes without being limited by a priori regional hypotheses. Future investigations should continue to quantify cortical thickness alterations related to alcohol and to relate cortical thickness to severity of alcoholism. Particularly, cortical thickness should be related to measures of alcohol consumption across a variety of drinkers to investigate whether quantity of alcohol consumed can predict brain changes. Further, relationships between cognitive dysfunction and cortical thickness measures in alcoholics should be explored.