If the depression noted following TBI was significantly determined by the neuro-anatomical effects of the brain injury as opposed to the sociocultural and clinical phenomena arising as a result of the injury (such as loss of self-esteem, pain, psychological reactions, etc.), then it would be anticipated that there would be a distinct pattern of impairment seen in depression in association with brain injury as opposed to depression independent of the injury. There would also be expected to be a strong association of clinical symptom with lesion site (Fleminger et al., 2003) as well as the demonstration of the "biologic gradient" (i.e., one would expect to see an increased risk of the affective disorders with increased severity of TBI).
Perhaps to begin with a discussion of the relationships between lesion site and postinjury depression it is best to indicate what the condition has not been associated with. Whilst the issue is by no means clear-cut, post-TBI depression has not been found to be related in any meaningful way to duration of loss of consciousness (Bornstein, Miller, & Van Schoor, 1989; Levin & Grossman, 1978), to the duration of PTA (Bornstein et al., 1989) or to the presence of skull fracture (Bornstein et al., 1989). Some relationships have been noted, however, with the level of neuropsychological sequelae (e.g., Bornstein et al., 1989; Dikmen & Reitan, 1977).
These findings are not too surprising in the context of the realization that we still do not have a very good understanding as to how the neuropsychological effects of the affective disorders may manifest in the absence of TBI. It is thus worthwhile to begin at the beginning and then see how this conundrum might be resolved.
As noted in the discussion of the phenomena associated with depression in the absence of TBI discussed above, research into the neurobiology of depression has been pursued at various levels of discourse (e.g., molecular, endocrine, neural, neuro-imaging, phenomenological), but as yet this enquiry has failed to provide a definitive account of the causal mechanism of the condition. One of the principal reasons for this is that both biological and psychological systems act in a series of feedback loops, making it difficult to identify which biochemical/behavioural/psychological/social changes preceded the others and thus eventually culminated in the pathology.
The "diathesis stress" model of depression (Nemeroff, 1998) provides an excellent means by which to integrate this evidence. This model proposes that individuals vary in terms of their sensitivity to stress-induced biological responses, due to their genetic endowment. Stressful experiences alter endocrine processes to the extent that normal homeostatic processes are compromised and depression ensues. A stressful experience may thus exhaust psychological coping by compromising underlying biological/motivation processes, while on the other hand (as recent evidence suggests), "stressed" glial cells may contribute to frontal lobe functional pathology via gradual atrophic changes (Drevets, 1999).
Duman, Heninger, and Nestler (1997) further develop this notion suggesting
One possibility is that many individuals who become depressed may have had a prior exposure to stress that causes a small amount of neuronal damage, but not enough to precipitate a behavioral change. If additional damage occurs as a result of normal aging or further stressful stimuli, these effects may then be manifested in the symptoms of mood disorder. These types of events could explain the decreased volume of specific brain structures in depression. (p. 604)
As discussed in chapter 3, it is clear that the frontal lobes play a crucial role in the perceptual interpretation of the emotional significance of external stimuli and a regulatory role over more primal (subcortical) brain structures such as the amygdala (Le Doux, 1998). Frontal lobe involvement in depression implicates the left and medial prefrontal cortex (PFC) (Elliott, Baker et al., 1997), the left dorsolateral PFC (Busch & Alpern, 1998), as well as the anterior cingulate (Devinsky, Morrell, & Vogt, 1995).
Abnormalities in brain structure and function have been consistently observed in patients with affective disorders and computed tomography (CT) scanning of these patients has revealed ventricular enlargement in bipolar disorders, unipolar depression, and mixed affective disorders (Post, 2000). Numerous studies have shown that elderly depressed individuals possess values more similar to those with irreversible dementia than to normals (Post, 2000). The best replicated finding across various depressed populations (young versus old, drug-naive and medication refractory disease and in patient subgroups) is a decrease in frontal lobe activity during restingstate functional imaging studies (Post, 2000). The changes involve the dorsolateral
(Brodmann's areas (BA) 9, 10, 46) as well as ventral and orbitofrontal cortex (BA 10, 11, 47). Most studies report bilateral changes although asymmetries have also been reported. In addition, limbic (amygdala), paralimbic (anterior temporal, cingulate), and subcortical (basal ganglia, thalamus) loci have been inconsistently identified.
Numerous positron emission tomography (PET) and single-photon emission computed tomography (SPECT) studies have demonstrated an inverse relationship between frontal activity and depression severity (Post, 2000) and significant correlations have been demonstrated with psychomotor speed, anxiety, and cognitive performance (e.g., Austin et al., 1992). The level of frontal hypometabolism noted is comparable in patients with depression of varying types (unipolar, bipolar, and with obsessive-compulsive disorder [OCD]) (Post, 2000) and there have been consistent findings of a negative correlations between HDRS and cerebral activity in depression (Post, 2000).
Mayberg's (1997, 2001, 2006) intriguing work proposes a working neuroana-tomical model of depression that proposes that depression comes about as a consequence of three neuroanatomical spheres of activity: (1) the dorsal compartment, which includes both neocortical and superior limbic components and is postulated to mediate cognitive aspects (apathy, psychomotor slowing, impaired attention, and executive functioning); (2) the ventral compartment including limbic, paralimbic, and subcortical regions, which mediate the circadian and vegetative components including sleep, appetite, libido, and endocrine disturbances; and (3) the rostral cingulate, which is distinct from both the dorsal and ventral components that may mediate interactions between the two. Support for the model comes from functional neuroimaging data as well as from event-related activity, electroencephalographic data, and neuropsychological measures. This earlier work is further supported by the very encouraging results arising from the application of deep brain stimulation of the subgenual cingulate area (Mayberg, 2006) in intractable depression.
Considerable dispute has raged in the earlier literature regarding lesion site and side and their effects on post-TBI depression. This arose from the early work of Pierre Flor Henry (e.g., 1974, 1976) who proposed that right hemisphere lesions culminated in affective disorders while left hemisphere lesions resulted in psychosis. In their influential early study of penetrating missile wounds following the Vietnam War discussed above, Grafman, Vance et al. (1986) reported that depressive symptoms were more commonly associated with penetrating missile wounds of the right hemisphere (particularly the right orbitofrontal cortex).
In his series of papers on 670 patients suffering from a variety of penetrating missile wounds Lishman (1966, 1968, 1978, 1997) noted that depression was also most commonly found with lesions of the right frontal lobe. This finding stands in contrast to subsequent studies on the poststroke depression literature (see Robinson, 2000 for a comprehensive review) that suggests that during the acute poststroke period, patients with left frontal lesions had a significantly higher frequency (i.e., 60% left frontal versus 0% right frontal; 13% left posterior versus 17% right posterior) of major depression than individuals with other lesion location.
More recent data has suggested that the left hemisphere is most likely implicated in post-TBI depression. Fedoroff et al. (1992) observed that in their 66 consecutive patients with CHI but no spinal cord or other organ system injury, interviewed one month postinjury using a structured clinical interview (PSE), an observer-rated symptom checklist (HDRS) and the mini-mental status examination (MMSE), 17 subjects (26%) met the DSM-III symptom and duration criteria for major depression and a further two patients had dysthymia. There was a significantly higher frequency of previous psychiatric disorder and alcohol abuse in the depressed subjects.
The presence of a left anterior lesions (i.e., left dorsolateral frontal and/or left basal ganglia lesions and, to a lesser degree, parietal-occipital and right hemisphere lesions) was the strongest correlate of those examined of major depression. These results lead the authors to speculate: "this finding suggests that the left dorsolat-eral frontal cortex and the left basal ganglia are critical structures in the left hemisphere as far as mood is concerned, and they may represent strategic locations for the initiation of major depression" (p. 922). They went on to suggest that this might be due to the lesions interrupting biogenic amine-containing neurons as they pass through the basal ganglia or frontal subcortical white matter.
While these views are interesting in themselves, it is perhaps more probative to have an oversight of the various studies undertaken in this complex issue to determine if any overarching trends emerge. In the comprehensive meta-analysis undertaken by Carson et al. (2000), 143 studies were identified that addressed this question and 48 of these met the criteria for inclusion. Of these 48 studies, 38 showed no association with lesion location, 2 showed increases with left-sided lesions, and 8 showed increases with right-sided lesions. The subsequent meta-analytic investigation indicated no evidence to support increased prevalence of depression with left as compared to right hemisphere strokes. The data also indicated no evidence for selective risk for depression following damage to the anterior regions of the brain (after taking into account that anterior strokes are the most common form of stroke).
There has, however, been some criticism of the review by Carson et al. (2000). The methodology of the review and, most notably the exclusion of some studies, has been a particular focus of concern (Dilley & Fleminger, 2006). A subsequent meta-analysis has suggested that the proximity of the lesion to the left frontal pole does predict depressive illness (Narushima, Kosier, & Robinson, 2003) and this observation has been further supported by a subsequent analysis from Finland (Vataja et al. 2004). Thus, while the possibility of a sided contribution to the emergence of depression is an interesting one, the veracity of this suggestion remains in contention.
One possible means of saving the hypothesis of a specific lesion prompting depression may be to make the assumption of "non-imaging detectable" neural damage as indicated by neuropsychological deficit. The lack of detailed injury data is a significant limitation, however, and lead Nelson, Drebing, Satz, and Uchiyama (1998) to observe: "More problematic is the inherent assumption that all CHI patients are alike. We know from the pathophysiology of head injury that different subtypes emerge involving, for example, orbitofrontal and dorsolateral areas of the brain" (p. 559).
While there are some doubts about the directness of the relationship between neu-roimaging evidence and functional evidence (e.g., Anderson et al., 1995), there is considerable debate in the literature as to whether, where, how, or why neuropsycho-logical deficits occur in depression (Caine, 1986; Elliott, Baker et al., 1997; Rohling et al., 2002). Various methodological and theoretical anomalies have prevented the emergence of any cohesive account of research findings thus far (Rosenstein, 1998).
As noted above, some authors (e.g., Reitan & Wolfson, 1997; Veiel, 1997) go so far as to suggest few, if any, intrinsic depressive deficits exist. Others, taking a more moderate view, concede that where deficits do occur, they are either "replete with inconsistencies" (Rosenstein, 1998, p. 139) or are very difficult to define. Dolan, Bench, Brown, Scott, and Frackowiak (1994) note "the underlying nature of depression related cognitive deficits has been a source of extensive theoretical debate. Psychological theories include motivational deficits, lack of effort, poor encoding strategies or defective processing resources" (p. 849).
McAllister-Williams, Ferrier, and Young (1998) also cogently observe "if neuro-psychological deficits were secondary to low mood, one would expect the degree of neuropsychological dysfunction to reflect this. Early studies support this but it has been challenged by recent studies" (p. 573).
Most authors agree that neuropsychological deficits are associated with depressive illness in some way or another. The central point of concern, as hinted above, is whether these deficits are "primary" (i.e., intrinsic to the condition) or "secondary" (i.e., the result of otherwise intact cognitive capacities failing due to lack of functional opportunity, such as via lowered subcortical arousal).
In both TBI and non-TBI affected individuals, depression exacerbates CHI-induced cognitive impairment (Brooks, 1984) and leads to poor social functioning (Busch & Alpern, 1998; Jorge, Robinson, Starkstein, & Arndt, 1994). Perhaps one of the best means of addressing these problems presented by the sometimes contradictory literature is to use meta-analysis.
One of the most useful studies along this line of enquiry is the analysis conducted by Hans Veiel (1997). Veiel's analysis only examined depressed young or middle aged individuals with clearly defined major depression. Using these exclusion criteria, he developed a database of 13 studies. He notes that the results of the studies can be classified into three levels of effect. He notes low levels of effect on studies that measure attention and concentration issues. He notes a moderate level of effect of depression on visuo-motor tracking, visual/spatial functions, and verbal fluency. He observes high levels of effect on measures of mental flexibility and control and composite indicators of brain impairment (i.e., the pathognomonic index of the Luria Nebraska Battery).
Veiel notes that 50% of patients with major depression will score two or more SDs below normals on the Trail Making Test—Part B (TMT[B]) or on the Colour-word form of the Stroop Task as well as on composite indicators (e.g., the pathognomonic indicator of Luria Nebraska Neuropsychological Battery [LNRB]), 15% will score in this range on tests of memory, visuomotor tracking/scanning, visual spatial functions and verbal fluency. On tests of simple attention, they will perform in a manner similar to normals. He notes that the level of disruption of cognition is of a severity comparable to that of severe TBI.
Was this article helpful?