Neuropsychological Impairment in Major Depression: Its Nature,Origin and Clinical Significance

Abstract

Neuropsychological impairment is well established as a feature of major depressive disorder (MDD) but studies have shown a variable pattern of impairment. This paper seeks first to clarify this by examining methodological and clinical factors that give rise to variability in study findings. Second, it examines theories of the origin of these neuropsychological abnormalities. Third, it reviews evidence regarding the clinical significance of different patterns of deficit. A selective review was undertaken of the literature with a particular emphasis on methodological factors, the influence of clinical subtypes and prevalent theories of neuropsychological abnormality. Methodological issues and the heterogeneity of MDD account for considerable variability in results. Specific investigation of the subtypes of psychotic MDD, melancholic MDD and bipolar depression reduces this heterogeneity and results are more consistent in the elderly. Hypothalamic–pituitary–adrenal axis dysfunction is associated with neuropsychological dysfunction in MDD although evidence of direct causation is not definitive at present. Impairment of executive and psychomotor function is a consistent finding, particularly in the elderly, and may reflect frontostriatal–limbic dysfunction. There is growing evidence that this may have clinical significance. It is suggested that future research take very careful account of the exact phenotype of MDD. Classification based on neuropsychological profile may, in fact, be useful. Further research should examine further the clinical importance of patterns of neuropsychological impairment.

Keywords

hippocampus HPA axis limbic system major depressive disorder neuropsychological function

Cognitive dysfunction has increasingly been recognized as a core feature of major depressive disorder (MDD) [1]. Over the past three decades many studies have investigated the pattern and magnitude of impairment during and between episodes, the neuropsychological domains affected, and the origin of these abnormalities. This review will focus specifically on neuropsychological function in MDD and its relationship to other aspects of the disorder. In view of the increasing literature on the neuropsychology of bipolar disorder and several recent reviews [2–4], we will focus on unipolar depression except where reference is made to studies including both unipolar and bipolar depression, or specifically comparing the two conditions.

We will first discuss some basic principles of neuropsychological testing then refer to methodological issues important in interpreting studies of neuropsychological testing in MDD. Second, because there is considerable variation in the phenotype of depression, we will not attempt to construct a typical neuropsychological profile of depression but will examine the clinical features that affect that profile. Third, we will examine possible explanations, both psychological and biological for the deficits seen in depression. Finally, we will examine the issue of whether impairments demonstrated represent a state or trait phenomenon and what their clinical implications may be.

Neuropsychological testing in MDD

There is much debate regarding the domains of function assessed by individual neuropsychological tests. Indeed, many neuropsychological tests involve a number of complex cognitive processes, and it has been suggested that assigning tests to specific cognitive categories is somewhat arbitrary [5]. For example, a single verbal memory task requires intact comprehension, attention, declarative and auditory memory and expressive language skills, making it difficult to attribute poor performance to a specific neuropsychological deficit [6].

The broad domains of functioning most commonly assessed in MDD include attentional processes, verbal and non-verbal learning and memory, and executive functions. Each of these is briefly discussed.

Attention

The construct ‘attention’ is central to neuropsychological testing although a clear and universally accepted definition of the term has yet to be agreed on [7]. Attention has been conceptualized as involving three components: selective attention, sustained attention, and divided attention [8]. Selective attention (commonly referred to as concentration) involves perception of incoming information while ignoring competing information. Sustained attention (commonly referred to as vigilance) refers to the capacity to maintain and focus on an activity over a period of time. Divided attention simply involves the ability to respond to more than one task at the same time. The attention system has limited capacity, although this varies between individuals and also within each individual at different times and in different contexts [7]. It is important to note that there are no pure measures of attention per se; performance can be measured only on particular tasks that are presumed to require high levels of attention. Traditional tests are usually global measures of attention, which assess an overall deficit in attention rather than analysing the individual components of attention. The latter can be examined in more specific computerized tasks but this has rarely been done in MDD to date [8].

Verbal and non-verbal learning and memory

Memory is a higher level cognitive function dependent on intact functioning of other cognitive processes. It is widely accepted that it is not a unitary concept, and numerous models have been developed to describe various memory systems.

One of the oldest distinctions is between working memory (sometimes referred to as short-term memory, although the terms are not strictly interchangeable; see [9]) and long-term memory. Working memory is involved in the storage and temporary maintenance of information. It is sometimes divided into visuospatial and auditory–verbal modalities. Long-term memories, on the other hand, are maintained from hours to years and have been consolidated or stored. The process by which this is thought to occur is termed long-term potentiation (see [10] for a brief review) and occurs predominantly in the hippocampus. Such memory is usually separated into two distinct systems: declarative (‘effortful’ or ‘cognitive’ learning) and non-declarative (‘habit-like’ learning, such as skills or procedures). Declarative memory is sometimes further subdivided into semantic memory (memory for facts) and episodic memory (memory for events and personal experiences; specifically autobiographical memory).

Specific impairments in memory are often discussed in terms of memory processes, such as acquisition or encoding, rehearsal, consolidation, and retrieval. These contribute to memory formation, but are not memory systems in themselves. It is also important to note that although particular tasks can be relatively specific to a particular memory system, several memory systems and processes may contribute to performance on any given test.

Executive function

Executive function has been described as ‘those capacities that enable a person to engage successfully in independent, purposive, self-serving behaviour’ [7] and include ‘higher-order’ cognitive processes, such as aspects of planning, judgement, decision-making, anticipation or reasoning, control of attention, and task management. Executive functions differ from specific cognitive functions, and deficits in executive functioning are more likely to be global with broader effects on behaviour. Some deficits in executive function are very apparent, such as impaired self-control, or difficulty in shifting attention. Others, however, are more subtle, such as decreased motivation and defects in planning and decision-making.

Historically, the terms ‘executive function’ and ‘frontal lobe function’ have been used interchangeably, largely because many patients with damage to the prefrontal cortex show deficits in planning, organization, decision making and other executive functions. However, a more recent view is that executive functions do not depend entirely on frontal lobe function, but instead are mediated by neural networks involving different regions with collaborative and overlapping functions [11, 12].

Methodological issues

Findings from the many studies of cognitive function in MDD have been variable, with some studies reporting relatively global cognitive deficits, some reporting specific deficits but in varying domains of cognitive function and other studies reporting no deficits at all. Some of this variability may be due to a number of methodological inconsistencies that need to be considered before further discussion. In fact, two early meta-analyses identified less than 25 studies as possessing sufficient methodological rigour to be particularly useful in the field [1, 5].

As will be discussed later in this review, many clinical factors may affect the neuropsychological profile of major depression. Relevant clinical features are the presence or absence of psychosis, melancholia, bipolar versus unipolar depression, comorbid personality disorder and comorbid anxiety. These factors must be adequately characterized in studies in order for the data to be interpreted.

Adequate matching of the experimental (depressed) group to the comparison (healthy) group is important. Essentially, the only difference likely to impact on cognitive function should be the presence of depression. Differences between groups that could give rise to a difference in neuropsychological function not attributable to depression include sex differences, phase of menstrual cycle in female subjects, age and pre-morbid cognitive function. We will review the issues of clinical factors and adequate matching in more detail in the following sections.

A wide variety of tasks has been used in investigation of neuropsychological function that vary in their psychometric properties and therefore their sensitivity to detect changes in cognitive function. A source of variance in the conclusions reached by studies lies in the fact that profiles are described solely in terms of their statistical significance rather than in the actual magnitude of the differences. Future studies should seek to expand on the presentation of results to include absolute differences and confidence intervals as well as assessment of effect size. The use of statistical techniques such as testing differences for evidence of true neuropsychological differential deficit should also be considered [13], as well as a more theoretical approach to the design of test batteries and their interpretation [14].

Effects of clinical factors on cognitive function in MDD

Variations in phenotype crucially affect the nature and extent of neuropsychological impairment in MDD. Therefore the best way to approach a characterization of neuropsychological impairment in MDD is to discuss the pattern in individual subtypes.

Severity, melancholia and inpatient status

The issue of what has variously been termed ‘endogenous, psychotic, vital, autonomous, endogenomorphic, Type A and melancholic depression’ continues to be debated. According to some research groups [15], melancholia is a distinct entity with a greater genetic component, particular biological abnormalities, better response to biological treatment, fewer psychosocial stressors and less evidence of personality dysfunction. Two major points are in dispute: first whether ‘melancholia’ is in fact a separate entity; and second, if it is, what the features are. The debate is summarized by Parker et al.: ‘To the binarians melancholia is a categorical entity; to the unitarians, depressive disorders vary dimensionally, with melancholia viewed merely as a more severe expression’ [16]. If a distinct, more biological form of depression can be characterized then it is in this form that more significant neuropsychological abnormalities would be expected to be apparent.

Studies investigating the relationship between melancholia and neuropsychological impairment have not shown a definitively different pattern unrelated to measures of severity. Austin et al. found that predominantly mnemonic tasks were impaired in a heterogeneous group of patients compared with controls, but that when a narrow definition of melancholia was used (CORE or Newcastle scales) to define a group of melancholic patients, executive dysfunction was found in the melancholic patents compared with controls [17]. However, there were no significant differences when the subtypes of depression were compared and severity used as a statistical covariate, perhaps suggesting that this was merely a function of melancholic patients tending to be more severely depressed.

Psychomotor slowing is an essential part of most definitions of melancholia and it has been argued that this is in fact the core feature of melancholia [18]. Not surprisingly then studies have suggested that patients with melancholia are more likely to demonstrate slowing on tasks measuring psychomotor speed. Pier et al. investigated whether psychomotor retardation was more cognitive or motor in nature by operationalizing distinct cognitive and motor variables for each neuropsychological task used [19]. They found that a group of medication-free melancholic inpatients showed both cognitive and motor slowing, whereas the non-melancholic depressed patients showed significant slowing only on the variables that reflected cognitive processes, suggesting that the motor slowing was specific to the melancholic patients.

Some studies have shown an association with severity of depression and selected aspects of neuropsychological performance although this is not consistent even between different depression rating scales. Elliott et al. found significant correlations between severity of depression and several tests of visuospatial learning and memory when using the Montgomery–Asberg Depression Rating Scale (MADRS), but not the Hamilton Depression Rating Scale (HDRS) [20]. Notable studies in very mildly depressed patients have also failed to show significant neuropsychological impairment. For example, Grant et al. examined patients with MDD, drug-free for 4 weeks, and healthy controls and observed a general absence of impairment in the patient sample [21]. Surprisingly, patients performed better than controls on several tasks, including verbal learning and memory, reaction time and match-to-sample latency. This is likely to be the result of the design of the study in which testing was performed following a 28 day medication-free period, during which subjects received supportive clinical management and single-blind placebos to ensure the presence of persistent depressive symptoms at the time of testing. This resulted in a very mild severity of illness in the patient group.

Studies that have examined the effects of hospitalization on neuropsychological performance have consistently demonstrated more severe impairment in inpatients than outpatients. Using the intradimensional/extradimensional (ID/ED) set-shifting task (a computerized test analogous to the Wisconsin Card-Sorting Test (WCST) involving executive function) in young depressed patients, Purcell et al. classified the 50% who completed the task as an ‘unimpaired group’ and the 50% who failed as an ‘impaired group’ [22]. The groups did not differ according to severity of depression, age of onset or length of illness, age, years of education, IQ or medication status. However, patients in the impaired subgroup were more likely to have reported a history of inpatient hospitalization compared with the unimpaired group. Similarly, Elliott et al. examined a cohort of middle-aged depressed patients, after separating them by current hospitalization [20]. The two groups did not differ in terms of age, IQ or severity, but the 11 inpatients were found to be significantly more impaired on computerized tests of delayed matching to sample, spatial span and response to failure than the 17 outpatients. It is especially interesting that this is found in both currently hospitalized patients and those with a past history of hospitalization [20, 22]. It is not possible to know exactly what it is about inpatients that gives rise to this phenomenon because inpatient status may arise from a number of different factors including melancholia, suicidality and episodes that are more severe, longer, recurrent or more treatment resistant [23]. Inpatients are also much more likely to have abnormalities of the hypothalamic–pituitary–adrenal (HPA) axis [24] and are more likely to have or have had psychotic episodes.

Psychotic depression

Other biological abnormalities appear to have an increased incidence in psychotic depression. For instance there appears to be a particularly high incidence of abnormalities of the HPA axis [25–27].

Studies comparing the performance of psychotic unipolar depressed patients with non-psychotic patients have consistently shown that the neuropsychological impairment in psychotic depression is significantly more widespread and severe. Those studies that have failed to find such differences have generally been in smaller sample sizes with a post-hoc division into those with and without psychotic features, rather than recruiting such groups from the outset. In a review and meta-analysis of the literature on neuropsychological impairment in psychotic depression, Fleming et al. noted similar methodological issues to those noted here [28]. Overall, however, there were differences across all domains of cognitive function between psychotic and non-psychotic depressed patients, the greatest differences being in verbal memory, executive functioning and psychomotor speed. Since that review, Gomez et al., comparing 29 psychotic with 24 non-psychotic depressed patients, found that patients with psychotic depression had more severe deficits in virtually every domain of functioning [29]. Specifically examining attention, Politis et al. found that psychotic depressed patients were significantly impaired on a variety of measures of selective and sustained attention compared to non-psychotic depressed patients, showing a similar level of impairment to a group of patients with schizophrenia [30]. Therefore there is good evidence for a more severe deficit in psychotic MDD than non-psychotic MDD but without good evidence of domain specificity.

Comparison between unipolar and bipolar depression

Comparisons of unipolar and bipolar depression have generally been small, with conflicting findings. For instance, Mojtabai et al. found no differences on any measure of attention, memory, executive function, and general intellectual functioning between MDD and bipolar depression [31]. Borkowska and Rybakowski found, however, that there was significantly greater impairment in bipolar depression than MDD across a range of executive functions as well as in general intellectual functioning compared with healthy controls [32]. The discrepancy is probably due to the confounding effect of other depressive subtypes because all subjects in the Mojtabai et al. study were psychotic. This highlights the importance of illness variables on performance and raises the question of whether the two disorders can ever be compared satisfactorily. For example, even if the groups could be matched for demographic profile and the severity of current episode, it would be impossible to equate overall illness characteristics such as medication, hospitalizations and previous affective episodes (of depression or mania).

Effects of age

Depression that occurs for the first time in later life is referred to as late-onset depression. This is usually defined according to an age of onset cut-off that varies across studies [33], but usually defines late onset as later than age 60. This distinction appears to have important implications for the neurobiological findings, in addition to neuropsychological abnormalities. This may be because of a significantly greater contribution of vascular pathology in the late-onset cases of depression [34]. Endogenous risk factors associated with late-onset depression include genetic factors, medical diseases, and cardiovascular disease, which may predispose to vascular lesions disrupting frontostriatal pathways or their modulating systems [35].

Age is associated with a progressive decline in neurocognitive functioning. It has been suggested that the specific domains that are affected are the same as those that are impaired in more severe (and melancholic-type) depression, and therefore that age must always be viewed as a significant confound [36]. In their meta-analysis, Christensen et al. reported a larger magnitude of effect of depression on neuropsychological function in patients over 60 years of age than in those under 60 years [37]. Individually, many studies have used age as a covariate in their statistical analyses or have examined the correlation between age and neuropsychological performance, although relatively few have directly compared the performance of older and younger depressed subjects versus groups of matched healthy controls.

Tarbuck and Paykel, and Lockwood et al. both examined the neuropsychological profile in groups of depressed patients below and above 60 years of age [38, 39]. The groups were otherwise well matched and were assessed on a comprehensive battery of tests. Overall, significant effects of age were observed on the majority of tests, particularly in psychomotor speed. Selected ‘complex’ or executive tasks were disproportionately impaired in the older depressed group. Differences between results in younger and older patients are therefore important, age should be taken into account in analysing data and this issue will be revisited frequently throughout this review.

Effects of comorbid personality disorder

MDD is commonly associated with a number of comorbid psychiatric disorders with an overrepresentation of cluster C and to a somewhat lesser degree cluster B personality disorders [40]. Most studies of cognitive function in MDD have not screened patients for the presence of personality disorders, despite the fact that borderline personality disorder particularly has been found to be associated with neuropsychological impairment. For this reason it is difficult to determine whether the profile of deficits is due to the depressive disorder or the co-occurring personality disorder. Only a few studies have systematically investigated the effects of MDD and comorbid personality disorders on neuropsychological performance, and conflicting findings have been reported. In one such study, Kurtz and Morey found that patients with borderline personality disorder and comorbid MDD showed poorer recall and recognition memory compared to healthy controls, and poorer recognition memory than patients with MDD only [41]. In contrast, however, Fertuck et al. found no differences on seven domains of cognitive function between groups of patients with borderline personality disorder and comorbid MDD, and patients with MDD only [42]. Bourke et al. showed a synergistic relationship between depression and borderline personality disorder with more marked impairment in executive function where both were present compared with either alone or neither [43]. The effect of comorbid personality disorder on cognitive function in depression should be investigated further and taken into account in future studies.

Effects of comorbid anxiety

MDD is often complicated by anxiety symptoms and actual comorbidity with anxiety disorders [44]. In spite of the high rate of comorbid anxiety disorders, many studies of cognitive function in MDD have failed to account for this. It may also be desirable to characterize the degree of anxiety symptoms that may have a significant bearing on neuropsychological profile. In a very large population study Kizilbash et al. classified subjects into high and low levels of anxiety and depression [45]. The combination of high levels of anxiety and depression gave rise to significant impairment of aspects of memory compared with high levels of either alone. There is little consistent research to date on the neuropsychological profile of individual anxiety disorders although some studies have found significant impairment in certain domains in specific disorders [46]. This then could interact with the profile seen in depression.

Diurnal variation of mood

Diurnal variation in mood commonly occurs in MDD. A recent study showed significant variation in 59% of patients, with 27% being worse in the morning and 32% worse in the evening [47]. However, relatively few studies have investigated the effects of diurnal variation on cognitive performance in MDD and not all have controlled time of testing. A feature of DSM-IV-defined melancholia is worse mood in the morning. A study of melancholic depressed patients found that when patients were tested in the morning, they showed marked impairment on a number of domains of function including attention and concentration/working memory, episodic memory, reaction time and speed of simultaneous match to sample [48]. In line with diurnal improvement in mood in the evening, the melancholic patients showed significant improvement on these tasks in the evening. In a similar study by Porterfield et al., impaired performance on a verbal fluency test was associated with diurnal variation of mood [49]. These findings have important implications for the timing of administration of cognitive tests, if the confounding effects of diurnal fluctuations in mood and cortisol levels on cognitive performance are to be minimized.

Effects of medication

The effects of antidepressant and other psychotropic medication and their withdrawal may be of particular importance in studies of neuropsychological function in less severely depressed or non-melancholic samples where impairment may be more subtle. It appears that antidepressants with anti-cholinergic properties impair aspects of neuropsychological function, while there may be less effect of selective serotonin re-uptake inhibitors or monoamine oxidase inhibitors. Many studies have attempted to control for antidepressant effects by using a wash-out period. Unfortunately, in some cases this has not been for long enough because although onset of antidepressant discontinuation symptoms is usually within several days of stopping an antidepressant, symptoms may last several weeks [50] and may affect cognitive function. Very few studies have examined patients who are drug-naïve or are even following a drug-free period of more than a few weeks. One study has examined patients drug free for at least 6 weeks prior to testing [51]. Indeed of 44 medication-free outpatients with moderate to severe MDD, 26 had never taken antidepressants. Compared with healthy controls, impairment was found on tests of sustained attention and executive function, as well as visuospatial memory. However, verbal declarative learning and memory, motor speed, and psychomotor functions were unimpaired. Clearly that study demonstrates that neuropsychological abnormalities in MDD are not wholly secondary to the effects of medication. A further recent study confirmed significant neuropsychological impairment in 38 drug-free depressed patients [52].

Effects of gender

Few studies of neuropsychological impairment in MDD have considered the effects of gender on cognitive performance despite clear evidence of differences in cognitive functioning between sexes [53]. Recent imaging studies also suggest that there may be important and complex differences in regional brain activation, volume and lateralization between men and women, particularly on emotion perception tasks. For instance, a meta-analysis of 65 neuroimaging studies of emotion found that male subjects showed more lateralization of emotional activity, and female subjects showed more brainstem activation in affective paradigms [54].

It is vitally important that studies control for phase of menstrual cycle in women, which has important effects on cognitive function [55].

Origin of neuropsychological impairment in MDD

Psychological theories have suggested that cognitive changes in MDD may be mediated by symptom-related factors. These include (i) reduced motivation on tasks involving effortful processing; (ii) catastrophic responses to failure; and (iii) mood-related interpretation and selective attention biases towards or away from disorder-related verbal or facial stimuli. Such theories do not exclude an underlying neurobiological abnormality as part of the cause. Neurobiological factors thought to be important include changes in functional neuroanatomy particularly in the frontostriatal–limbic system, structural change such as hippocampal atrophy and hormonal changes particularly in the HPA axis.

Effortful versus automatic processing

It has been suggested that effort is the principal factor mediating neuropsychological changes in MDD, and a recent meta-analysis provided evidence that performance on tasks requiring effortful processing is more impaired than performance on more automatic tasks. Minimal effect sizes were found on tests of primary, semantic and working memory (relatively automatic processes), with the most prominent effects being in memory processes, such as the effortful encoding of information and inefficiency of retrieving poorly encoded information from declarative memory [1]. Many studies of neuropsychological impairment in MDD have found deficits on tests of attention and executive function, both domains typically requiring effortful processing. However, the effortful processing hypothesis does not explain why deficits are often reported on non-effortful tasks. For instance, Porter et al. found not only ‘high-effort’ tasks to be impaired in MDD, but also relatively automatic processes, such as recognition memory in both visual and verbal domains, despite intact verbal declarative memory [51].

A related theory is that cognitive deficits in MDD may be due to reduced motivation [56]. This is closely related to the effortful processing hypothesis, but reduced motivation does not explain the lack of impairment in certain cognitive domains and the specificity of deficits seen in others if it is the major factor in mediating neuropsychological dysfunction.

Catastrophic response to failure

Several studies have investigated patients’ responses to perceived failure on cognitive tasks in MDD. In elderly patients, Beats et al. described a catastrophic response to perceived failure on a computerized version of the Tower of London task. They found that once patients made an error, subsequent performance deteriorated markedly [57]. Elliott et al. reported a different response to failure on the delayed matching to sample task and the Tower of London Task [20]. In controls, but not depressed patients, the likelihood of succeeding on subsequent trials was increased following feedback that the wrong response had been given. However, some subsequent studies have failed to replicate these findings, possibly because of the exact nature of the task or feedback or the statistical procedures used [58, 59].

Mood-related attentional and memory biases in information processing

Many studies have shown differences in attentional biases and interpretation of emotional information (usually words or faces) between patients with MDD and healthy controls. The Stroop colour–word test has been extensively used to investigate executive function deficits. Emotional variants of the Stroop task have been developed for different psychiatric disorders, using emotionally valenced words relevant to the disorder. The increased interference during colour naming of the emotionally valenced words is thought to be the result of an inability to inhibit emotionally salient material. Several studies have demonstrated performance deficits in MDD on emotional Stroop tasks. For instance, Segal et al. administered a task requiring subjects to name the colour in which positive and negative adjectives, differing in the degree to which they described the subject, were presented [60]. The target adjectives were primed by emotional phrases that also varied according to degree of self-reference. Results indicated that depressed patients were slower to colour-name self-descriptive negative targets primed by self-descriptive negative phrases compared with healthy controls. This effect was not found for positive targets with depressed patients, and the controls showed no effect of prime–target relation for material in either valence.

Studies have also investigated mood-congruent memory biases in MDD, but methodological factors have contributed to conflicting findings. For instance, studies that have found mood-congruent memory biases have tended to use self-referent emotional information as stimuli, whereas those that have not found biases have tended to use more general depression-related emotional words [61]. Obviously, personally related material is generally more easily remembered. Other studies have utilized affective go/no-go tasks to investigate affective set shifting, a component of executive function. In a study by Murphy et al., depressed patients showed an affective bias towards words with negative emotional valence while manic patients exhibited the opposite bias, towards words with a positive emotional valence [62].

There is increasing interest in abnormal recognition of facial emotion in MDD. Some studies have shown a negative bias in the interpretation of emotions when sad, ambiguous, or neutral facial expressions have been presented [63], whereas others have shown reduced positive attentional biases [64]. Other studies have reported either specific or generalized deficits in the perception of various emotions, which differ according to the clinical characteristics of the patient groups or the testing paradigms and stimuli used [65]. More recently, studies of regional brain function have begun to reveal the relationships between these impairments and the specific brain structures and neural pathways involved in emotion perception in MDD [66].

Recently it has been suggested that antidepressants may therefore work in a manner similar to that of psychological treatments that aim to redress negative biases in information processing, but by doing this in a pharmacological way [67]. This fits with evidence regarding the effects of various pharmacological manipulations on emotional processing [68, 69].

Frontostriatal–limbic abnormalities

Neuroimaging studies have implicated limbic–thalamic–cortical circuits in the neuropathology and neurophysiology of depression [70], with involvement of the amygdala, medial thalamus, orbital and medial prefrontal cortices, and the limbic–cortical–striatal–pallidal–thalamic circuits. The neuroanatomical locus of executive functions has long been believed to be the frontal lobes, but it is now understood that such processes activate frontal–subcortical loops. Dysfunction of this circuitry has been hypothesized to be central to many of the observed neuropsychological impairments in MDD [71, 72]. Alexopoulos et al. have also proposed a relationship between MDD in old age and frontostriatal–limbic dysfunction [35]. This is supported by the presence of white matter hyperintensities in subcortical structures in elderly with MDD and executive dysfunction [73].

Functional imaging studies add support to this hypothesis. Drevets has demonstrated abnormalities in activation of the prefrontal and limbic regions in melancholic depression [74], together with abnormal metabolism in caudate nucleus and frontal regions [70]. Functional neuroimaging has shown activation of the anterior cingulate (part of the frontostriatal–limbic pathways) during performance of the traditional Stroop task in healthy subjects [75, 76]. Performance of this task been shown to be abnormal in several studies of MDD and altered anterior cingulate activity has also been shown during Stroop performance [77].

Abnormalities of the HPA axis and hippocampal structure and function

Various abnormalities of the HPA axis have been reported in MDD and may manifest as increased cortisol secretion (see [78] for a recent review). It has been suggested that neuropsychological impairment in MDD is directly related to excess cortisol secretion [79]. Several studies in the elderly show a direct relationship between abnormal dexamethasone suppression (impaired negative feedback inhibition of cortisol secretion) and impaired mnemonic function [80, 81]. In younger populations the evidence is not so clear. One study supports a correlation between abnormal feedback and mnemonic function [82], and Rubinow et al. showed a correlation between cortisol secretion (24 h urinary cortisol) and performance on the Halsted Category Test, a general test of cognitive function [83]. Egeland et al. found that high levels of morning salivary cortisol were associated with executive impairments and memory storage and retrieval deficits in patients with recurrent MDD [84] but in a larger study, of moderately depressed outpatients, there was no evidence of correlation between measures of cortisol and cognition [51] but neither was there evidence of hypercortisolaemia. The relationship may therefore be relevant only in subtypes of MDD with HPA axis dysregulation. Specifically in psychotic depression, rates of HPA axis abnormality are high (see the ‘Psychotic depression’ section) and one study has suggested a relationship between cortisol secretion and various aspects of cognitive dysfunction [29]. Bipolar depression has also been associated with higher rates of HPA axis abnormality [85] and there is preliminary evidence of a relationship between abnormal dexamethasone suppression and working memory in patients with bipolar disorder (although not specifically depression) [86]. It may therefore be that in subtypes of MDD in which there is significant HPA axis abnormality, this is a factor that impairs cognitive function.

There is ongoing debate about the exact nature of the neuropsychological impairment induced by excess of corticosteroids. Several studies have investigated the effects of administration of corticosteroids to human volunteers. Some have reported a specific reduction in verbal declarative memory function [87], which may be the result of a specific deficit in memory retrieval [88, 89]. However, there is also evidence to suggest that working memory function may be more sensitive than declarative memory to the effects of elevated corticosteroid levels [90, 91]. This was supported in a recent treatment study targeting HPA axis abnormalities in bipolar depression using the glucocorticoid-receptor antagonist mifepristone (RU-486); improvements in spatial working memory were observed with no change in verbal declarative memory [92].

Increased levels of corticosteroids also have structural effects on the brain, particularly the hippocampus. Brown et al. reported that patients on long-term prescription corticosteroid therapy had smaller hippocampal volumes, lower scores on neuropsychological tests of executive function and declarative memory and higher ratings of depressive and general psychiatric symptoms [93]. Starkman et al. used MRI to examine patients with Cushing's disease before and after treatment and found that hippocampal formation volume increased by up to 10% after treatment [94].

In depression, however, the relationship between brain structure, cortisol levels and memory impairment may be more complex. Overall, studies of hippocampal volume in depression show a reduction compared with healthy subjects and this reduction appears to be related to the number of episodes of MDD [95, 96]. However, in adult subjects (age 18–60 years) Vythilingam et al. found no difference in hippocampal volume between depressed patients and controls, despite evidence of specific neuropsychological impairments in the former [97]. Baseline plasma or urinary free cortisol (UFC) was not related to either hippocampal volume or memory deficits and successful treatment with antidepressants did not change hippocampal volume but did improve memory function and reduce UFC excretion. O'Brien et al. carried out neuropsychological assessment, salivary cortisol profiling and MRI in patients over age 60 and controls at baseline and at 6 month follow up [98]. Depressed subjects showed multiple neuropsychological impairments and had hypercortisolaemia and a reduction in right hippocampal volume. However, hippocampal volume reduction was not associated with increased cortisol levels but was significantly correlated with continuing memory deficits at 6 months. Persisting ‘mild neuropsychological impairment’ was seen in 41% subjects at 6 months and was associated with reduced hippocampal volume but not severity of depression or cortisol levels. Vascular factors may begin to become important in the elderly depressed and may obscure a clear relationship between hippocampal structure and HPA axis function [98]. The hippocampus has an important role in learning and memory but is also part of the limbic system and of frontostriatal–limbic circuits and may therefore influence executive function [99]. In patients with either a previous or current episode of depression, Frodl et al. found reduced hippocampal volume which correlated with reduced performance on the WCST [100], a task of executive function involving frontostriatal–limbic circuitry [101].

State versus trait

The extent to which deficits in neuropsychological function in depression are purely state related, or whether they represent trait markers of the illness remains an important area of research interest. While there is now little doubt that neuropsychological impairment can be observed during episodes, studies have now shown that specific deficits persist long into recovery and may therefore represent trait markers. It is also possible that impairment in certain domains is a function of the depressive state while in other domains it is a function of a permanent underlying neurobiological abnormality, that is, a state marker.

In a follow-up study of severely depressed inpatients, Trichard et al. reported that selective executive impairments (Stroop performance) were found to persist, even after clinical remission [102]. Similarly euthymic outpatients with a history of recurrent depression have also been found to exhibit specific executive functioning deficits in the absence of declarative memory impairment [103]. However, other studies have reported more extensive deficit that may relate to a more chronic course of illness. Specifically in patients with a history of recurrent melancholic depression, Marcos et al. found impairments both in executive function and in visuospatial memory [104]. In non-symptomatic patients with a history of chronic depression there was also a broader profile of impairment, including domains of immediate memory and attention, executive function and (psycho)motor speed [105]. Weiland-Fiedler et al. compared remitted MDD drug-free patients with controls and found deficits in aspects of executive function and sustained attention, although many of these were not significant when residual symptoms were controlled for statistically [106]. No evidence of verbal or visuospatial learning and memory impairment was found. However, more recently Paelecke-Habermann et al. have reported more widespread attentional and executive impairments in 40 remitted (defined as a period of at least 3 months) MDD patients compared to controls [107]. Taken together these findings would seem to suggest that memory deficits are the most closely linked to depression severity and therefore clinical state. The findings for executive/attentional impairments are more equivocal; although some specific aspects may represent trait markers (which remains to be established), it is possible that this may be a consequence of such tests being more sensitive and adversely affected even by residual, subsyndromal depression. Recent twin studies have also supported the notion that genetic liability to affective disorder is associated with neuropsychological impairment but they do not show a selective impairment. Christensen et al. found that healthy twins discordant for unipolar disorder showed lower performance on almost all measures of cognitive function examined: selective and sustained attention, executive function, language processing and working and declarative memory, and also after adjustment for demographic variables, subclinical symptoms and minor psychopathology [108]. Effects were also found in twins discordant for bipolar disorder although the profile was far more specific and limited to episodic and working memory.

Data regarding residual impairment are clearer in the elderly in whom there is consistent evidence of residual executive impairment and reduced psychomotor speed following recovery [57, 109]. A recent study, however, also demonstrated significant memory impairment and interestingly, that 23% of patients who were defined as cognitively normal at baseline were impaired 1 year later. The issue of predisposition to dementia in subjects with MDD is one of increasing interest and research [110].

Clinical significance

There is increasing evidence that measuring neuropsychological function may be a useful tool in predicting clinical response, predicting relapse and the general course of the illness and perhaps even as a sensitive measure of response to treatment [111].

Prediction of response

Early studies examining the impact of neuropsychological impairments on response to treatment were relatively small or used a variety of treatments [112–115]. However, a recent study investigated response to citalopram 40 mg for 8 weeks in 112 elderly patients with MDD. Impairment of aspects of executive function measured by the Stroop and the ‘initiation and perseveration’ part of the Mattis dementia rating scale [116] predicted poor response to treatment [117]. Several studies using other methodologies suggest that frontostriatal–limbic dysfunction is present in groups of patients with poor outcome. Subcortical white matter hyperintensities are associated with executive dysfunction [118] and with poor outcome in MDD in the elderly [119, 120].

Recently, in a small preliminary study we have reported the follow-up data from our earlier study [51]. Of the 25 MDD patients who returned at follow up, 11 were defined as remitted and 14 as not remitted [121]. Significantly less baseline psychomotor dysfunction was observed in patients who remitted compared to those who did not, while analysis of the neuropsychological change scores between assessments revealed a significantly greater improvement in verbal memory in patients who remitted compared with those who did not [121].

Prediction of long-term prognosis

It has been suggested that these residual cognitive deficits may be one of the greatest barriers to rehabilitation [122]. In 58 elderly patients remitted from MDD, Alexopoulos et al. found that impairment of executive function but not memory deficit predicted relapse [123]. However, in a similar study, Butters et al. found no such association between baseline neuropsychological function and relapse [124]. We know of no similar studies in younger people.

Summary

Over the past three decades many studies have demonstrated neuropsychological impairment in MDD. The pattern and extent of deficits ranges from no apparent dysfunction to severe motor and cognitive dysfunction, and it is likely that methodological factors such as variability in clinical subtype, severity, age, medication, and testing paradigms may account for some of the inconsistencies in findings. Separation of the subtypes of melancholic depression, psychotic depression and bipolar depression may reduce this heterogeneity but still does not produce consistent results. It is possible that in MDD, specific neuropsychological profiles themselves represent useful constructs and that, for instance, MDD with frontostriatal–limbic impairment represents a useful subtyping with predictive value and implications for treatment, both pharmacological and psychological. Construction of such neuropsychological endophenotypes is itself, however, hampered by tests with considerable interindividual variability and a lack of specificity for particular brain circuitry.

The evidence to date suggests that impairment in executive and psychomotor function in MDD, probably reflecting dysfunction in frontostriatal–limbic circuitry, is more prevalent in the elderly, melancholic/severe, psychotic and those with bipolar disorder. Particularly in these groups it may be seen as a trait marker and indicate a poorer prognosis. In the elderly it may be related to subcortical vascular lesions. Memory dysfunction seems more related to the state of depression.

Dysregulation of the HPA axis as measured by the dexamethasone suppression test correlates well with neuropsychological impairment in many studies of MDD in the elderly. This is not necessarily causal – both may simply indicate more severe biological dysfunction. However, hippocampal atrophy in more chronic cases of MDD may be related to hypercortisolaemia. Hypercortisolaemia and hippocampal dysfunction may produce executive as well as memory dysfunction. Treatments directly altering HPA axis dysfunction may be useful in improving cognitive function in MDD.

Finally, change in neuropsychological function may be a useful measure of treatment response that could be used effectively both in clinical trials and even clinical practice to determine resolution of core abnormalities and guide prognosis. It is important also to realize that even subtle cognitive dysfunction may adversely affect individuals with MDD, causing significant disruption to social and occupational functioning.

References

1. Zakzanis

Leach

Kaplan

. On the nature and pattern of neurocognitive function in major depressive disorder. Neuropsychiatry Neuropsychol Behav Neurol 1998; 11: 111–119.

2. Robinson

Ferrier

. Evolution of cognitive impairment in bipolar disorder: a systematic review of cross-sectional evidence. Bipolar Disord 2006; 8: 103–116.

3. Savitz

Solms

Ramesar

. Neuropsychological dysfunction in bipolar affective disorder: a critical opinion. Bipolar Disord 2005; 7: 216–235.

4. Robinson

Thompson

Gallagher

. A meta-analysis of cognitive deficits in euthymic patients with bipolar disorder. J Affect Disord 2006; 93: 105–115.

5. Veiel

. A preliminary profile of neuropsychological deficits associated with major depression. J Clin Exp Neuropsychol 1997; 19: 587–603.

6. Keefe

. The contribution of neuropsychology to psychiatry. Am J Psychiatry 1995; 152: 6–15.

7. Lezak

. Neurophysiological assessment3rd edn. Oxford University Press, New York 1995.

8. Mialet

Pope

Yurgelun-Todd

. Impaired attention in depressive states: a non-specific deficit?. Psychol Med 1996; 26: 1009–1020.

9. Shah

, Miyake

. Models of working memory: an introduction. In:. Miyake

, Shah

, eds., Models of working memory: mechanisms of active maintenance and executive control. Cambridge: Cambridge University Press, 1999:1–27.

10.

10. Reid

Stewart

. Seizures, memory and synaptic plasticity. Seizure 1997; 6: 351–359.

11.

11. Alexander

DeLong

Strick

. Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annu Rev Neurosci 1986; 9: 357–381.

12.

12. Elliott

. Executive functions and their disorders. Br Med Bull 2003; 65: 49–59.

13.

13. Crawford

Blackmore

Lamb

Simpson

. Is there a differential deficit in fronto-executive functioning in Huntington's Disease?. Clin Neuropsychol Assess 2000; 1: 3–19.

14.

14. Thompson

Hamilton

Gray

. Executive and visuospatial sketchpad resources in euthymic bipolar disorder: implications for visuospatial working memory architecture. Memory 2006; 14: 437–451.

15.

15. Carney

MWP

Roth

Garside

. The diagnosis of depressive syndromes and the prediction of E.C.T. response. Br J Psychiatry 1965; 111: 659–674.

16.

16. Parker

Hadzi-Pavlovic

Boyce

. Issues in classification: II. Classifying melancholia. Melancholia: a disorder of movement and mood, Parker

Hadzi-Pavlovic

. Cambridge University Press, Cambridge 1996; 20–37.

17.

17. Austin

Mitchell

Wilhelm

. Cognitive function in depression: a distinct pattern of frontal impairment in melancholia?. Psychol Med 1999; 29: 73–85.

18.

18. Parker

Hadzi-Pavlovic

. Development and structure of the CORE system. Melancholia: a disorder of movement and mood, Parker

Hadzi-Pavlovic

. Cambridge University Press, Cambridge 1996; 82–129.

19.

19. Pier

Hulstijn

Sabbe

. Differential patterns of psychomotor functioning in unmedicated melancholic and nonmelancholic depressed patients. J Psychiatr Res 2004; 38: 425–435.

20.

20. Elliott

Sahakian

McKay

Herrod

Robbins

Paykel

. Neuropsychological impairments in unipolar depression: the influence of perceived failure on subsequent performance. Psychol Med 1996; 26: 975–989.

21.

21. Grant

Thase

Sweeney

. Cognitive disturbance in outpatient depressed younger adults: evidence of modest impairment. Biol Psychiatry 2001; 50: 35–43.

22.

22. Purcell

Maruff

Kyrios

Pantelis

. Neuropsychological function in young patients with unipolar major depression. Psychol Med 1997; 27: 1277–1285.

23.

23. Stage

Bech

Gram

Kragh-Sorensen

Rosenberg

Ohrberg

. Are in-patient depressives more often of the melancholic subtype? Danish University Antidepressant Group. Acta Psychiatr Scand 1998; 98: 432–436.

24.

24. Maes

Calabrese

Meltzer

. The relevance of the inversus outpatient status for studies on HPA-axis in depression: spontaneous hypercortisolism is a feature of major depressed inpatients and not of major depression per se. Prog Neuropsychopharmacol Biol Psychiatry 1994; 18: 503–517.

25.

25. Nelson

Davis

. DST studies in psychotic depression: a meta analysis. Am J Psychiatry 1997; 154: 1497–1503.

26.

26. Posener

DeBattista

Williams

Chmura Kraemer

Kalehzan

Schatzberg

. 24-hour monitoring of cortisol and corticotropin secretion in psychotic and nonpsychotic major depression. Arch Gen Psychiatry 2000; 57: 755–760.

27.

27. Keller

, Flores

, Gomez

. Cortisol circadian rhythm alterations in psychotic major depression. Biol Psychiatry 2006; 60:275–281.

28.

28. Fleming

Blasey

Schatzberg

. Neuropsychological correlates of psychotic features in major depressive disorders: a review and meta-analysis. J Psychiatr Res 2004; 38: 27–35.

29.

29. Gomez

Fleming

Keller

. The neuropsychological profile of psychotic major depression and its relation to cortisol. Biol Psychiatry 2006; 60: 472–478.

30.

30. Politis

Lykouras

Mourtzouchou

Christodoulou

. Attentional disturbances in patients with unipolar psychotic depression: a selective and sustained attention study. Compr Psychiatry 2004; 45: 452–459.

31.

31. Mojtabai

Bromet

Harvey

Carlson

Craig

Fennig

. Neuropsychological differences between first-admission schizophrenia and psychotic affective disorders. Am J Psychiatry 2000; 157: 1453–1460.

32.

32. Borkowska

Rybakowski

. Neuropsychological frontal lobe tests indicate that bipolar depressed patients are more impaired than unipolar. Bipolar Disord 2001; 3: 88–94.

33.

33. Krishnan

. Biological risk factors in late life depression. Biol Psychiatry 2002; 52: 185–192.

34.

34. Baldwin

O'Brien

. Vascular basis of late-onset depressive disorder. Br J Psychiatry 2002; 180: 157–160.

35.

35. Alexopoulos

Buckwalter

Olin

Martinez

Wainscott

Krishnan

. Comorbidity of late life depression: an opportunity for research on mechanisms and treatment. Biol Psychiatry 2002; 52: 543–558.

36.

36. Austin

Mitchell

Goodwin

. Cognitive deficits in depression: possible implications for functional neuropathology. Br J Psychiatry 2001; 178: 200–206.

37.

37. Christensen

Griffiths

Mackinnon

Jacomb

. A quantitative review of cognitive deficits in depression and Alzheimer-type dementia. J Int Neuropsychol Soc 1997; 3: 631–651.

38.

38. Tarbuck

Paykel

. Effects of major depression on the cognitive function of younger and older subjects. Psychol Med 1995; 25: 285–295.

39.

39. Lockwood

Alexopoulos

Van Gorp

. Executive dysfunction in geriatric depression. Am J Psychiatry 2002; 159: 1119–1126.

40.

40. Parker

Roussos

Austin

Hadzi-Pavlovic

Wilhelm

Mitchell

. Disordered personality style: higher rates in non-melancholic compared to melancholic depression. J Affect Disord 1998; 47: 131–140.

41.

41. Kurtz

Morey

. Verbal memory dysfunction in depressed outpatients with and without borderline personality disorder. J Psychopathol Behav Assess 1999; 21: 141–156.

42.

42. Fertuck

Marsano-Jozefowicz

Stanley

. The impact of borderline personality disorder and anxiety on neuropsychological performance in major depression. J Pers Disord 2006; 20: 55–70.

43.

43. Bourke

Porter

Sullivan

. Neuropsychological function in bulimia with comorbid borderline personality disorder and depression. Acta Neuropsychiatr 2006; 18: 162–167.

44.

44. Young

Abelson

Cameron

. Effect of comorbid anxiety disorders on the hypothalamic-pituitary-adrenal axis response to a social stressor in major depression. Biol Psychiatry 2004; 56: 113–120.

45.

45. Kizilbash

Vanderploeg

Curtiss

. The effects of depression and anxiety on memory performance. Arch Clin Neuropsychol 2002; 17: 57–67.

46.

46. Nitschke

Heller

. Distinguishing neural substrates of heterogeneity among anxiety disorders. Int Rev Neurobiol 2005; 67: 1–42.

47.

47. Joyce

Porter

Mulder

. Reversed diurnal variation in depression: associations with a differential antidepressant response, tryptophan: large neutral amino acid ratio and serotonin transporter polymorphisms. Psychol Med 2005; 35: 511–517.

48.

48. Moffoot

Bennie

. Diurnal variation of mood and neuropsychological function in major depression with melancholia. J Affect Disord 1994; 32: 257–269.

49.

49. Porterfield

Cook

Deary

Ebmeier

. Neuropsychological function and diurnal variation in depression. J Clin Exp Neuropsychol 1997; 19: 906–913.

50.

50. Coupland

Bell

Potokar

. Serotonin reuptake inhibitor withdrawal. J Clin Psychopharmacol 1996; 16: 356–362.

51.

51. Porter

Gallagher

Thompson

Young

. Neurocognitive impairment in drug-free patients with major depressive disorder. Br J Psychiatry 2003; 182: 214–220.

52.

52. Gualtieri

Johnson

Benedict

. Neurocognition in depression: patients on and off medication versus healthy comparison subjects. J Neuropsychiatry Clin Neurosci 2006; 18: 217–225.

53.

53. Postma

Jager

Kessels

Koppeschaar

van Honk

. Sex differences for selective forms of spatial memory. Brain Cognition 2004; 54: 24–34.

54.

54. Wager

Phan

Liberzon

Taylor

. Valence, gender, and lateralization of functional brain anatomy in emotion: a meta-analysis of findings from neuroimaging. Neuroimage 2003; 19: 513–531.

55.

55. Symonds

Gallagher

Thompson

Young

. Effects of the menstrual cycle on mood, neurocognitive and neuroendocrine function in healthy premenopausal women. Psychol Med 2004; 34: 93–102.

56.

56. Schmand

Kuipers

Van der Gaag

Bosveld

Bulthuis

Jellema

. Cognitive disorders and negative symptoms as correlates of motivational deficits in psychotic patients. Psychol Med 1994; 24: 869–884.

57.

57. Beats

Sahakian

Levy

. Cognitive performance in tests sensitive to frontal lobe dysfunction in the elderly depressed. Psychol Med 1996; 26: 591–603.

58.

58. Shah

O'Carroll

Rogers

Moffoot

Ebmeier

. Abnormal response to negative feedback in depression. Psychol Med 1999; 29: 63–72.

59.

59. Murphy

Robbins

MTW

Sahakian

. Neurophysiological impairment in patients with major depressive disorder: the effects of feedback on task performance. Psychol Med 2003; 33: 455–467.

60.

60. Segal

Gemar

Truchon

Guirguis

Horowitz

. A priming methodology for studying self-representation in major depressive disorder. J Abnorm Psychol 1995; 104: 205–213.

61.

61. Ellwart

Rinck

Becker

. Selective memory and memory deficits in depressed inpatients. Depression Anxiety 2003; 17: 197–206.

62.

62. Murphy

Sahakian

Rubinsztein

. Emotional bias and inhibitory control processes in mania and depression. Psychol Med 1999; 29: 1307–1321.

63.

63. Hale

. Judgment of facial expressions and depression persistence. Psychiatry Res 1998; 80: 265–274.

64.

64. Surguladze

Young

Senior

Brebion

Travis

Phillips

. Recognition accuracy and response bias to happy and sad facial expressions in patients with major depression. Neuropsychology 2004; 18: 212–218.

65.

65. Rubinow

Post

. Impaired recognition of affect in facial expression in depressed patients. Biol Psychiatry 1992; 31: 947–953.

66.

66. Lawrence

Williams

Surguladze

. Subcortical and ventral prefrontal cortical neural responses to facial expressions distinguish patients with bipolar disorder and major depression. Biol Psychiatry 2004; 55: 578–587.

67.

67. Harmer

Hill

Taylor

Cowen

Goodwin

. Toward a neuropsychological theory of antidepressant drug action: increase in positive emotional bias after potentiation of norepinephrine activity. Am J Psychiatry 2003; 160: 990–992.

68.

68. Harmer

Perrett

Cowen

Goodwin

. Administration of the beta-adrenoceptor blocker propranolol impairs the processing of facial expressions of sadness. Psychopharmacology 2001; 154: 383–389.

69.

69. Harmer

Bhagwagar

Perrett

Vollm

Cowen

Goodwin

. Acute SSRI administration affects the processing of social cues in healthy volunteers. Neuropsychopharmacology 2003; 28: 148–152.

70.

70. Drevets

. Neuroimaging studies of mood disorders. Biol Psychiatry 2000; 48: 813–829.

71.

71. Goodwin

. Neuropsychological and neuroimaging evidence for the involvement of the frontal lobes in depression. J Psychopharmacol 1997; 11: 115–122.

72.

72. Rogers

Bradshaw

Pantelis

Phillips

. Frontostriatal deficits in unipolar major depression. Brain Res Bull 1998; 47: 297–310.

73.

73. Krishnan

KRR

Hays

Blazer

. MRI-defined vascular depression. Am J Psychiatry 1997; 154: 497–501.

74.

74. Drevets

. Functional neuroimaging studies of depression: the anatomy of melancholia. Annu Rev Med 1998; 49: 341–361.

75.

75. MacDonald

Cohen

Stenger

Carter

. Dissociating the role of the dorsolateral prefrontal and anterior cingulate cortex in cognitive control. Science 2000; 288: 1835–1838.

76.

76. Pardo

Pardo

Janer

Raichle

. The anterior cingulate cortex mediates processing selection in the Stroop attentional conflict paradigm. Proc Natl Acad Sci USA 1990; 87: 256–259.

77.

77. Liotti

Mayberg

. The role of functional neuroimaging in the neuropsychology of depression. J Clin Exp Neuropsychol 2001; 23: 121–136.

78.

78. Porter

Gallagher

. Abnormalities of the HPA axis in affective disorders: clinical subtypes and potential treatments. Acta Neuropsychiatr 2006; 18: 193–209.

79.

79. McAllister-Williams

Ferrier

Young

. Mood and neuropsychological function in depression: the role of corticosteroids and serotonin. Psychol Med 1998; 28: 573–584.

80.

80. O'Brien

Schweitzer

Ames

Tuckwell

Mastwyk

. Cortisol suppression by dexamethasone in the healthy elderly: effects of age, dexamethasone levels, and cognitive function. Biol Psychiatry 1994; 36: 389–394.

81.

81. Mitchell

. The contribution of hypercortisolaemia to the cognitive decline of geriatric depression. Int J Geriatr Psychiatry 1995; 10: 401–409.

82.

82. Wauthy

Ansseau

von Frenckell

Mormont

Legros

. Memory disturbances and dexamethasone suppression test in major depression. Biol Psychiatry 1991; 30: 736–738.

83.

83. Rubinow

Post

Savard

Gold

. Cortisol hypersecretion and cognitive impairment in depression. Arch Gen Psychiatry 1984; 41: 279–283.

84.

84. Egeland

Lund

Landro

. Cortisol level predicts executive and memory function in depression, symptom level predicts psychomotor speed. Acta Psychiatr Scand 2005; 112: 434–441.

85.

85. Rush

Giles

Schlesser

. The dexamethasone suppression test in patients with mood disorders. J Clin Psychiatry 1996; 57: 470–484.

86.

86. Watson

Thompson

Ritchie

Nicol Ferrier

Young

. Neuropsychological impairment in bipolar disorder: the relationship with glucocorticoid receptor function. Bipolar Disord 2006; 8: 85–90.

87.

87. Newcomer

Craft

Hershey

Askins

Bardgett

. Glucocorticoid-induced impairment in declarative memory performance in adult humans. J Neurosci 1994; 14: 2047–2053.

88.

88. de Quervain

Henke

Aerni

. Glucocorticoid-induced impairment of declarative memory retrieval is associated with reduced blood flow in the medial temporal lobe. Eur J Neurosci 2003; 17: 1296–1302.

89.

89. de Quervain

Roozendaal

Nitsch

McGaugh

Hock

. Acute cortisone administration impairs retrieval of long-term declarative memory in humans. Nat Neurosci 2000; 3: 313–314.

90.

90. Young

Sahakian

Robbins

Cowen

. The effects of chronic administration of hydrocortisone on cognitive function in normal male volunteers. Psychopharmacology 1999; 145: 260–266.

91.

91. Lupien

Gillin

Hauger

. Working memory is more sensitive than declarative memory to the acute effects of corticosteroids: a dose-response study in humans. Behav Neurosci 1999; 113: 420–430.

92.

92. Young

Gallagher

Watson

Del-Estal

Owen

Nicol Ferrier

. Improvements in neurocognitive function and mood following adjunctive treatment with mifepristone (RU-486) in bipolar disorder. Neuropsychopharmacology 2004; 29: 1538–1545.

93.

93. Brown

Woolston

Frol

. Hippocampal volume, spectroscopy, cognition, and mood in patients receiving corticosteroid therapy. Biol Psychiatry 2004; 55: 538–545.

94.

94. Starkman

Giordani

Gebarski

Berent

Schork

Schteingart

. Decrease in cortisol reverses human hippocampal atrophy following treatment of Cushing's disease. Biol Psychiatry 1999; 46: 1595–1602.

95.

95. Campbell

Marriott

Nahmias

MacQueen

. Lower hippocampal volume in patients suffering from depression: a meta-analysis. Am J Psychiatry 2004; 161: 598–607.

96.

96. Videbech

Ravnkilde

. Hippocampal volume and depression: a meta-analysis of MRI studies. Am J Psychiatry 2004; 161: 1957–1966.

97.

97. Vythilingam

Vermetten

Anderson

. Hippocampal volume, memory, and cortisol status in major depressive disorder: effects of treatment. Biol Psychiatry 2004; 56: 101–112.

98.

98. O'Brien

Lloyd

McKeith

Gholkar

Ferrier

. A longitudinal study of hippocampal volume, cortisol levels, and cognition in older depressed subjects. Am J Psychiatry 2004; 161: 2081–2090.

99.

99. Wall

Messier

. The hippocampal formation–orbitomedial prefrontal cortex circuit in the attentional control of active memory. Behav Brain Res 2001; 127: 99–117.

100.

100. Frodl

Schaub

Banac

. Reduced hippocampal volume correlates with executive dysfunctioning in major depression. J Psychiatry Neurosci 2006; 31: 316–323.

101.

101. Ottowitz

Dougherty

Savage

. The neural network basis for abnormalities of attention and executive function in major depressive disorder: implications for application of the medical disease model to psychiatric disorders. Harvard Rev Psychiatry 2002; 10: 86–99.

102.

102. Trichard

Martinot

Alagille

. Time course of prefrontal lobe dysfunction in severely depressed inpatients; a longitudinal study. Psychol Med 1995; 25: 79–85.

103.

103. Clark

Sarna

Goodwin

. Impairment of executive function but not memory in first-degree relatives of patients with bipolar I disorder and in euthymic patients with unipolar depression. Am J Psychiatry 2005; 162: 1980–1982.

104.

104. Marcos

Salamero

Gutierrez

Catalan

Gasto

Lazaro

. Cognitive dysfunctions in recovered melancholic patients. J Affect Disord 1994; 32: 133–137.

105.

105. Paradiso

Lamberty

Garvey

Robinson

. Cognitive impairment in the euthymic phase of chronic unipolar depression. J Nerv Ment Dis 1997; 185: 748–754.

106.

106. Weiland-Fiedler

Erickson

Waldeck

. Evidence for continuing neuropsychological impairments in depression. J Affect Disord 2004; 82: 253–258.

107.

107. Paelecke-Habermann

Pohl

Leplow

. Attention and executive functions in remitted major depression patients. J Affect Disord 2005; 89: 125–135.

108.

108. Christensen

Kyvik

Kessing

. Cognitive function in unaffected twins discordant for affective disorder. Psychol Med 2006; 36: 1119–1129.

109.

109. Abas

Sahakian

Lewy

. Neuropsychological deficits and CT scan changes in elderly depressives. Psychol Med 1990; 20: 507–520.

110.

110. Steffens

Welsh-Bohmer

Burke

. Methodology and preliminary results from the neurocognitive outcomes of depression in the elderly study. J Geriatr Psychiatry Neurol 2004; 17: 202–211.

111.

111. Venn

Watson

Gallagher

Young

. Facial expression perception: an objective outcome measure for treatment studies in mood disorders?. Int J Neuropsychopharmacol 2005; 9: 1–17.

112.

112. Dunkin

Leuchter

Cook

Kasl-Godley

Abrams

Rosenberg-Thompson

. Executive dysfunction predicts nonresponse to fluoxetine in major depression. J Affect Disord 2000; 60: 13–23.

113.

113. Simpson

Baldwin

Jackson

Burns

. Is subcortical disease associated with a poor response to antidepressants? Neurological, neuropsychological and neuroradiological findings in late-life depression. Psychol Med 1998; 28: 1015–1026.

114.

114. Kalayam

Alexopoulos

. Prefrontal dysfunction and treatment response in geriatric depression. Arch Gen Psychiatry 1999; 56: 713–718.

115.

115. Potter

Kittinger

Wagner

Steffens

Krishnan

. Prefrontal neuropsychological predictors of treatment remission in late-life depression. Neuropsychopharmacology 2004; 29: 2266–2271.

116.

116. Mattis

. Dementia Rating Scale. Psychological Assessment Resources, Odessa, Florida 1989.

117.

117. Alexopoulos

Kiosses

Heo

Murphy

Shanmugham

Gunning-Dixon

. Executive dysfunction and the course of geriatric depression. Biol Psychiatry 2005; 58: 204–210.

118.

118. Aizenstein

Nebes

Meltzer

. The relation of white matter hyperintensities to implicit learning in healthy older adults. Int J Geriatr Psychiatry 2002; 17: 664–669.

119.

119. Hickie

Scott

Mitchell

Wilhelm

Austin

Bennett

. Subcortical hyperintensities on magnetic resonance imaging: clinical correlates and prognostic significance in patients with severe depression. Biol Psychiatry 1995; 37: 151–160.

120.

120. O'Brien

Ames

Chiu

Schweitzer

Desmond

Tress

. Severe deep white matter lesions and outcome in elderly patients with major depressive disorder: follow up study. BMJ 1998; 317: 982–984.

121.

121. Gallagher

, Robinson

, Gray

, Porter

, Young

. Neurocognitive function following remission in major depressive disorder: a potential objective marker of response?. Aust N Z J Psychiatry (in press).

122.

122. Chamberlain

Sahakian

. Cognition in mania and depression: psychological models and clinical implications. Curr Psychiatry Rep 2004; 6: 451–458.

123.

123. Alexopoulos

Meyers

Young

. Executive dysfunction and long-term outcomes of geriatric depression. Arch Gen Psychiatry 2000; 57: 285–290.

124.

124. Butters

Bhalla

Mulsant

. Executive functioning, illness course, and relapse/recurrence in continuation and maintenance treatment of late-life depression: is there a relationship?. Am J Geriatr Psychiatry 2004; 12: 387–394.