Assessment of emotion






Catherine HARMER,DPhil
University Department of Psychiatry
Warneford Hospital
Oxford, UK

Assessment of emotion


by C. Harmer, United Kingdom



Emotional dysfunction is a critical feature of disorders such as depression and anxiety, but it can be difficult to fully quantify and explore using clinical rating scales alone. In recent years, there has been significant progress in the development of cognitive paradigms to tap into different aspects of emotional processing across the domains of attention, interpretation, and memory. These approaches have been applied to characterize both the cognitive neuropsychology and the role of emotional dysfunction in depression and anxiety and to elucidate pharmacological and psychological treatment action. This article reviews different approaches for assessing emotion in healthy volunteers and in patient groups using cognitive paradigms in behavioral and neuroimaging models. These studies have revealed consistent and partly dissociable effects of depression and anxiety on emotional processing measures. Furthermore, these emotional processing markers are targeted early following administration of antidepressant and anxiolytic drug treatments. Such effects have been seen in the absence of changes in subjective experience, suggesting that they may be more sensitive measures to index emotional bias and response. This approach is a useful strategy to understand depression and anxiety and provides an experimental medicine model to test out hypotheses of treatment action and to evaluate novel compounds in development for disorders involving emotional dysfunction.

Medicographia. 2013;35:337-343 (see French abstract on page 343)



Disorders such as major depression and anxiety involve dysfunction of various aspects of emotional response and regulation. These disorders are typically diagnosed by clinical interview involving assessment of subjective experiences (for example, low mood or anhedonia). A variety of measurements exist to aid the clinician in diagnosing, monitoring, or quantifying levels of depression and anxiety. However, different clinical scales are often used in different contexts (such as primary compared with secondary care, or with psychological compared with pharmacological treatments) and may tap into slightly different aspects of depression.1 There is increasing interest in utilizing objective measures of emotional response measured with cognitive paradigms which might also be less resistant to reporting biases or difficulty in identifying or talking about one’s own emotional experiences. Such an approach may help us understand the mechanisms underlying emotional dysfunction and its treatment.2 This review will focus on the use of these cognitive paradigms to tap into different aspects of emotional processing and response and also the questionnaire and rating scale methods which may usefully complement these assessments.

Cognitive paradigms

In our day-to-day life, we are exposed to a myriad of social and emotional cues, which are often ambiguous and can be viewed from different perspectives. How we respond to this kind of emotional information is affected by what information we attend to in the first place, how we perceive or interpret this information, and what we remember later. Consistent with this idea, there is now broad experimental evidence that these different cognitive domains are affected in emotional disorders such as depression.3,4 Such evidence is consistent with cognitive models which propose that negative schema (or knowledge structures) in depression are maintained by negative biases in emotional processing, which together fuel the depression cycle.5 In particular, incoming information is filtered so that stimuli or events in line with the depression schema are overrepresented, leading to increased inflow and memory of negative over positive items. A similar approach has been suggested for anxiety, with information being oversampled for threat-relevant cues, thereby promoting excessive reactivity to potential threat. The existence of these biases in depression and anxiety has been widely described and characterized and has led to the development of cognitive paradigms which can tap into these different aspects of emotional processing.3,4 Such paradigms have the potential not only to inform us about the mechanisms underlying emotional disorders, but also to help us understand how different treatment approaches work to reverse this kind of emotional dysfunction.

_ Attention to emotional information
A variety of experimental paradigms can be used to assess different aspects of attention to emotional stimuli and this review will focus on two of the most common methods: the emotional Stroop task and the dot-probe paradigm. The emotional Stroop test is a measure of interference produced by emotional content on an unrelated response. This is a variant of the classic color-naming Stroop task where participants are asked to report the color of the font in which a word is presented, but ignore the written word itself. When the color and written word are incongruent (ie, red font for the written word “yellow”), interference effects give rise to slower responses. In a similar vein, the emotional Stroop task utilizes word stimuli with an emotional valence to interfere with the ability to make a speeded response (eg, naming the color of the word “death”). Studies using the emotional Stroop task have demonstrated that anxiety disorders are associated with longer reaction times when naming the color of threatening words compared with neutral or positive words.6 Attentional biases toward emotional stimuli have also been reported in depression, although this interference effect does not seem to be restricted to depression-relevant stimuli (see meta-analysis7). The emotional Stroop task also has a number of methodological caveats, which can make interpretation of the behavioral findings complex. Rather than reflecting attentional capture, it is possible that the delayed naming of the emotional words could reflect cognitive avoidance of the stimuli.8 In addition, interference may arise from a more generalized emotional arousal in response to the threatening or negative words, which could lead to a delay or inhibition in response selection.9





An alternative paradigm which has been extensively used is the dot-probe task (Figure 1).10 In this task, two stimuli (typically words or images) are displayed simultaneously on a screen, in two separate locations. One of these stimuli usually has an emotional value, whereas the other one is neutral. After a brief period, the words or images disappear and a probe (for example, one or two dots) appears on the screen, either in the place of the emotional stimulus or in the place of the neutral stimulus. The volunteers are asked to indicate the position or type of probe as quickly as possible. The premise behind this task is that if attention favors the emotional stimulus, relative reaction time to detect the probe will be faster when it replaces the emotional compared with the neutral stimulus. The results from this task are less easily explained by general arousal or bias effects and it can provide a “snapshot” of attentional allocation by altering the stimulus exposure duration to enable dissection of effects on attentional capture versus disengagement.

Results from the dot-probe task suggest that anxiety is associated with relatively faster responses to probes that replace threatening stimuli than to probes that replace neutral stimuli. Again, this is suggestive of increased attentional vigilance to the location of a threatening cue (see review11). As with the emotional Stroop task, such attentional biases toward threat have been demonstrated even when the stimuli are presented subliminally, suggesting that they may be operating at a relatively automatic level of processing.12 The dot-probe task is also sensitive to attentional negative bias in depression,13 which may be at least partly distinct from that seen in anxiety. While attentional bias is apparent in depression at relatively long stimuli durations (eg, 500 ms-1000 ms13), there are potential differences when the stimuli are presented for shorter durations. For example, Mogg et al reported an increased shift of attention toward negative words presented subliminally in anxious, but not depressed participants, whereas the opposite pattern was seen with supraliminally presented stimuli.12 These findings have led to the suggestion that anxiety may be associated with increased orienting to threat, whereas depression could represent a problem in disengaging from negative stimuli.11


Figure 1
Figure 1. An example dot-probe paradigm.

Volunteers are asked to fixate centrally and two stimuli are presented. One of
these stimuli is replaced by a probe to which the volunteer has to respond (eg,
by indicating if one or two dots have been presented). If attention has been allocated
to the negative stimulus, then relative reaction time will be faster if the
probe is in the same location as this stimulus (as in this example), compared
with its being presented in the opposite location. This allows vigilance to positive
and negative stimuli to be computed.



Studies using pharmacological challenges in the dot-probe task of the Emotional Test Battery (ETB) have revealed effects which are consistent with this framework. Administration of the selective serotonin reuptake inhibitor (SSRI) citalopram increased attentional vigilance toward positive stimuli and/or away from negative stimuli in healthy volunteers compared with double-blind administration of placebo.14,15 We have hypothesized that such effects relate to action of SSRIs on anxiety-relevant processing, and consistent with this, the anxiolytic diazepam also reduced attention to negative compared with positive information in this task.16 Such effects of diazepam and citalopram were evident at relatively short exposure durations, consistent with a role for early attentional processes in anxiety and drug action.

_ Approach and inhibition
Information processing bias can also be assessed with the affective go–no go task (a measure included in the Cambridge Neuropsychological Test Automated Battery [CANTAB]). In this task, participants are presented with a series of words which are either positive (eg, joyful), negative (eg, hopeless), or neutral (eg, element). In a given block, participants are asked to respond as quickly as possible to a given affective category and inhibit responses to the other categories, therefore providing a measure of approach, inhibition, and switching. Negative bias in this task has been shown in adult and adolescent depression, seen, for example, as quicker responses to sad versus happy stimuli,17,18 and a similar pattern is emulated following depletion of the amino acid precursor of serotonin, tryptophan, in healthy volunteers.19 Such effects suggest that approach and inhibition in this paradigm may model processing biases in depression.

_ Facial expression perception
Our ability to identify emotional states of others from rapid emotional expressions is a key aspect of social cognition which is affected in various psychiatric disorders. Depression has been associated with negative biases in the interpretation of facial expressions characterized as a relative inability to detect positive facial expressions, such as happiness and/or increased sensitivity to facial expressions displaying negative cues, such as sadness or fear.2 This kind of bias has been found to predict depression levels 3 and 6 months later20 and subsequent relapse to depression,21 consistent with a key role for perceptual biases in the maintenance of this disorder. Facial expression recognition can be measured by different tasks, but recent versions have capitalized on advances in computerized graphic manipulation techniques to create stimuli with different intensity levels.22 This morphing technique involves blending two prototype photographs of the same individual in different proportions to create a continuum between neutral expression and each emotion or between easily confused emotions, such as happiness and surprise or disgust and anger.

In the ETB, different examples and intensity levels of six basic emotions (anger, disgust, fear, surprise, happy, and sadness) are presented.2 Each face has been blended between the prototype emotion and neutral expression in 10% steps, which are then presented in a randomized order. Participants are asked to classify the emotional expression of each face using a labeled response box, allowing the measurement of recognition accuracy, speed of correct responses, and misclassifications for each emotion. This task is sensitive to depression23 and to antidepressant drug administration.23,24 In particular, antidepressants such as the SSRI citalopram were found to decrease the perception of negative facial expressions (including anger, disgust, fear, and sadness) in healthy volunteers.24 Such effects would be expected to reverse negative biases seen in depression and reduce the impact of this key maintaining factor in this disorder. Indeed, early change in perception of facial expressions of emotion is related to the emergence of therapeutic response seen over time.2

Along with other measures included in the ETB, this method has been applied to characterize novel drugs in development for depression, providing information about effects in human models, clinical profile, and dose.2 Using this approach, it was assessed whether the novel antidepressant agomelatine, which acts as a melatonergic agonist and 5-hydroxytryptamine receptor 2C (5-HT2C) antagonist, would also affect emotional processing despite its different pharmacological profile to conventional treatments. Consistent with this proposal, 7 days’ administration of agomelatine at 25 mg specifically decreased the recognition of sad facial expressions compared with placebo in healthy volunteers.25 This effect on sadness was more selective than the more generalized effects found with drugs like citalopram or reboxetine in this task,24 suggesting a more targeted reversal of depression-relevant processing bias following agomelatine. The use of this cognitive neuropsychological approach applied to drug development may therefore allow the dissection of different processes involved in drug action and the generation of hypotheses about antidepressant potential and application.


Figure 2
Figure 2. The emotion-potentiated startle task.

Bursts of loud white noise are delivered through the headphones and the eyeblink startle response
is measured with electromyography. The emotion-potentiated startle effect is seen as increased
amplitude of startle in response to the loud noise when viewing unpleasant compared with neutral
or pleasant stimuli. Please note that to retain integrity of the affective stimuli used in this kind of
paradigm, fictitious example stimuli are displayed.
Top: Image of female performing emotion-potentiated startle task. © The Author. Left: Clay pot on
light background. © bryljaev/123RF. Middle: Flying Balloons. © Alexander Fediachov/123RF.
Right: Image of female pointing gun at somebody breaking and entering. © Justin Kral/123RF.



_ Emotional memory
Memory bias in depression has been well characterized, with negative stimuli being disproportionately remembered in shortand long-term memory tests in depressed patients compared with matched controls.3,4 Memory tasks used in depression typically involve explicit recall of emotional verbal material, although other more implicit measures of memory have been used, including facial affective priming26 and anagram solving.27 Self-referent stimuli may be particularly susceptible to negative bias in depression and many studies have used personality adjectives as stimuli (for example, words such as honest, original, or mean). In the ETB, a first stage of encoding is used (emotional categorization), where participants are presented with personality adjectives and asked whether they themselves would like or dislike to be referred to with each characteristic. This is a variant on an original task which asked volunteers whether each characteristic described them or not (“me” or “not me”; for example28). This modification was made to ensure relatively similar response choice across groups to allow both reaction time and memory to be explored, unconfounded by differences in endorsement. This task is sensitive to negative bias in depressed patients,23 volunteers at high risk of depression,29 and dysphoric participants (unpublished data).

Effects on emotional memory may also help to dissociate those processes relevant to depression versus anxiety. In particular, the negative bias seen with explicit memory tasks in depression is not consistently found in anxiety disorders or in volunteers with high versus low trait scores on anxiety measures.27,30 Emotional memory also seems to be the domain most consistently influenced by antidepressant drug treatment. Hence, acute reboxetine, mirtazapine, and duloxetine, and repeated administration of agomelatine, reboxetine, and citalopram all increased the recall of positive versus negative stimuli in this task in healthy volunteers.25,31-33

Similarly, a single dose of reboxetine was found to reverse negative bias in memory in depressed patients.23 Consistent with the distinction between depression and anxiety on emotional memory bias, purer anxiolytics such as diazepam do not typically affect performance on this measure, despite having action on other tasks in the battery related to anxiety (such as the dotprobe task and startle responses16). Such a profile suggests again that by understanding the cognitive neuropsychology of drug action, we can start to predict treatment effects and clinical profiles to optimize randomized clinical trial assessments of novel drugs in development.

_ The emotion-potentiated startle task
The emotion-potentiated startle task (EPST) is a human analog of the fear-potentiated startle task used to screen for anxiolytic agents in preclinical studies. There are a number of variations in the way in which this task can be administered, but all involve electromyographic (EMG) measurement of eye blink amplitude following a loud (eg, 90dB) burst of noise (Figure 2). The affective component of this task is typically induced through presentation of emotive picture stimuli or through expectation of an electric shock. Increased startle reactivity has been described in anxiety and is increased following social stress.34 In pharmacological studies, anxiolytic drug treatments have been reported to decrease emotion-potentiated startle responses in healthy volunteer models.2 Thus, acute administration of diazepam16 and mirtazapine31 and 7 days’ treatment with citalopram24 decreased startle responses in this paradigm. Agomelatine administration also led to abolition of the emotion-potentiated startle responses in healthy volunteers, suggestive of an anxiolytic profile.25

_ Use of functional imaging models
The effects of depression and anxiety and their treatment on emotional processing tasks provide important information about functional changes that may be involved in etiology and drug response. These measures can also be complemented by neuroimaging investigations, based on similar emotional processing tasks, which can reveal underlying neural mechanisms involved in these emotional changes. The neural circuitry of negative bias in depression is believed to involve interactions between the amygdala, hippocampus, anterior cingulate cortex, and dorsolateral prefrontal cortex (PFC).3 For example, increased drive in the amygdala in depression would be expected to increase responses to and interest in negative stimuli and facilitate preferential encoding into memory through projections to the hippocampus.3 Self-referent memory tasks also tap into areas involved in self-processing, such as medial PFC and precuneus, while the emotional Stroop test reliably activates the anterior cingulate cortex, which is involved in conflict detection and monitoring.2 These measures can all be adapted to work well in a functional imaging context, but the probe used most consistently across studies involves presentations of facial expressions of emotion.

Studies using this approach in functional magnetic resonance imaging (fMRI) have revealed increased amygdala reactivity to fear and/or sad facial expressions in depression, which is normalized following antidepressant treatment.35-37 These effects are seen prior to changes in mood or other symptoms of depression, consistent with the data from behavioral models reviewed above. Decreased amygdala response to negative versus positive cues can also be observed in healthy volunteer models and across different antidepressant drug classes.2 Further research is needed to isolate those processes, which are relevant to antidepressant versus anxiolytic properties of these drug treatments. Indeed, similar effects have also been described in the treatment of anxiety disorders. For example, a recent study by Phan et al (2012) revealed that patients with generalized social phobia also showed increased amygdala responses to threatening facial expressions, but decreased ventromedial PFC (vmPFC) responses to the same stimuli.38 The vmPFC plays a key role in the regulation of emotional reactivity, presumably via its connections with the amygdala, and this pattern is therefore consistent with impaired function of this regulatory network.39 Of particular relevance, this imbalance in activity was normalized following a 12-week SSRI treatment.38

The use of fMRI in these investigations therefore has the potential to uncover key mechanisms involved in emotional dysfunction and treatment response. However, it provides a different profile of advantages and disadvantages to behavioral testing alone. Ideally, fMRI should be complemented by behavioral results to allow any changes in neural response to be interpreted in light of evidence for impaired or affected processing in direct measures of performance. In addition, when planning a pharmacological fMRI study, it is important to be mindful of possible drug-induced changes in hemodynamic response or neural coupling, which can confound the blood oxygen- level–dependent (BOLD) outcome measure.40 It is therefore important to build in appropriate control conditions and tasks to assess changes in hemodynamic response unrelated to the task of interest or to supplement fMRI assessment with more direct measures of neural activity.40 However, despite these limitations, the use of these neuroimaging methods has the potential to uncover key processes and mechanisms for which behavioral results are inconclusive. fMRI can therefore enhance behavioral measures of emotional processing both in our understanding of illness and its treatment and in application to drug development and screening of novel agents for depression and anxiety.

_ Mood and subjective experience rating scales
Measures of emotional bias are complemented by a thorough examination of mood and subjective experience, based on self-report or clinician-rated scales. To assist in diagnosis and provide a measure of illness severity, the clinician-rated 17- item Hamilton Depression Rating Scale (HAM-D41) and the Montgomery-Asberg Depression Rating Scale (MADRS42) have been well validated for use in clinical trials and research studies. The self-report measures such as the Beck Depression Inventory (BDI)43 and the Patient Health Questionnaire-944 also provide highly relevant information, with the latter being used as a diagnosis aid in primary care settings. However, it should be noted that although each scale provides an overall measure of depression, the emphasis on different symptoms seen in this disorder is different.1 For example, the HAM-D has a greater number of items relating to sleep and anxiety than the MADRS or the BDI, whereas the BDI places more emphasis on pessimism and feeling of guilt. It is therefore important to consider the different symptom clusters in depression which might be relevant to a dissection of the cognitive neuropsychological maintaining factors.

In addition to rating scales of depression and anxiety, which typically ask about symptoms and experiences over a period of one or two weeks, it is also necessary to have measures of current mood state which may be more responsive to pharmacological induced change in healthy volunteer groups. Again, a variety of measures exist to monitor different aspects of everyday subjective state. The positive and negative affective schedules provide information across a period of time set by the experimenter (ie, over the last week or day or hour) with two subscales detailing positive and negative experiences.45 The Befindlichkeits Scale (BFS) is widely used in conjunction with the ETB to provide a measure of change in mood and energy levels.46 This requires the volunteer to choose between two adjectives (eg, shy or bold, sluggish or animated) describing current state and therefore may be more sensitive to minor changes in state than symptom-based measures. Anxiety can be assessed by the Spielberger State-Trait Anxiety Inventory (STAI), which provides two measures, a measure of phasic (or state) anxiety response in a given situation and a trait measure of more stable patterns of anxiety across time.47 The STAI has been used across studies to recruit volunteers with high or low levels of trait anxiety and to index the response to stressors and pharmacological manipulations. Finally, visual analog scales are used widely in the field of pharmacological challenges to monitor side effects and key effects on mood and anxiety.48 These involve the presentation of a line (typically 10 cm) with a description at the top (eg, alert) and labels at each end (running from “Not at all” to “Extremely”). Volunteers are asked to mark a place on the line which represents how they feel in relation to the descriptor. These can be used repeatedly within the same experimental setting and therefore provide a more finely grained analysis of subjective state, which can be applied in healthy volunteers and patient groups.

Conclusions

Emotional processing can be measured using objective measures of cognitive function both in behavioral and neuroimaging models. These have been used to explore the underlying etiology and treatment of emotional disorders, such as depression and anxiety. Emotional processing is affected earlier in treatment than changes in subjective experience and these early effects are predictive of clinical action occurring later in time. This approach may therefore be more sensitive to change than measures of subjective state and provide an early marker of response for treatment studies. Objective performance on cognitive tests is also less likely to be affected by reporting bias or difficulty of accessing relatively implicit cognitive- emotional states. The application of this kind of approach to complement diagnosis of emotional disorders, in addition to treatment response, represents an exciting possibility for future studies. Together with clinical and nonclinical assessment of mood, anxiety, and experience, this approach can be used to reveal key aspects of emotional dysfunction and to understand how established and candidate treatments may work to treat these disabling conditions. _


References
1. Uher R, Farmer A, Maier W, Rietschel M, et al. Measuring depression: comparison and integration of three scales in the GENDEP study. Psychol Med. 2008; 38:289-300.
2. Harmer CJ, Cowen PJ, Goodwin GM. Efficacy markers in depression. J Psychopharmacol. 2011;25:1148-1158.
3. Disner SG, Beevers CG, Haigh EA, Beck AT. Neural mechanisms of the cognitive model of depression. Nat Rev Neurosci. 2011;12:467-477.
4. Roiser JP, Elliott R, Sahakian BJ. Cognitive mechanisms of treatment in depression. Neuropsychopharmacology. 2012;37:117-136.
5. Beck AT, Rush AJ, Shaw BF, Emery G. Cognitive Therapy of Depression. New York, NY: Guilford; 1979.
6. Williams JM, Mathews A, MacLeod C. The emotional Stroop task and psychopathology. Psychol Bull. 1996;120:3-24.
7. Epp AM, Dobson KS, Dozois DJ, Frewen PA. A systematic meta-analysis of the Stroop task in depression. Clin Psychol Rev. 2012;32:316-328.
8. Wells A, Matthews G. Attention and Emotion: A Clinical Perspective. Hove, UK: Lawrence Erlbaum Associates; 1994.
9. Algom D, Chajut E, Lev S. A rational look at the emotional Stroop phenomenon: a generic slowdown, not a Stroop effect. J Exp Psychol Gen. 2004;133: 323-338.
10. MacLeod C, Mathews A, Tata P. Attentional bias in emotional disorders. J Abnorm Psychol. 1986;95:15-22.
11. Mogg K, Bradley BP. A cognitive-motivational analysis of anxiety. Behav Res Ther. 1998;36:809-848.
12. Mogg K, Bradley BP, Williams R. Attentional bias in anxiety and depression: the role of awareness. Br J Clin Psychol. 1995;34:17-36.
13. Peckham AD, McHugh RK, Otto MW. A meta-analysis of the magnitude of biased attention in depression. Depress Anxiety. 2010;27:1135-1142.
14. Browning M, Reid C, Cowen PJ, Goodwin GM, Harmer CJ. A single dose of citalopram increases fear recognition in healthy subjects. J Psychopharmacol. 2007;21:684-690.
15. Murphy SE, Yiend J, Lester KJ, Cowen PJ, Harmer CJ. Short-term serotonergic but not noradrenergic antidepressant administration reduces attentional vigilance to threat in healthy volunteers. Int J Neuropsychopharmacol. 2009;12: 169-179.
16. Murphy SE, Downham C, Cowen PJ, Harmer CJ. Direct effects of diazepam on emotional processing in healthy volunteers. Psychopharmacology (Berl). 2008; 199:503-513.
17. Murphy FC, Sahakian BJ, Rubinsztein JS, et al. Emotional bias and inhibitory control processes in mania and depression. Psychol Med. 1999;29:1307-1321.
18. Kyte ZA, Goodyer IM, Sahakian BJ. Selected executive skills in adolescents with recent first episode major depression. J Child Psychol Psychiatry. 2005;46: 995-1005.
19. Murphy FC, Smith KA, Cowen PJ, Robbins TW, Sahakian BJ. The effects of tryptophan depletion on cognitive and affective processing in healthy volunteers. Psychopharmacology (Berl). 2002;163:42-53.
20. HaleWW3rd. Judgment of facial expressions and depression persistence. Psychiatry Res. 1998;80:265-274.
21. Bouhuys AL, Geerts E, Gordijn MC. Depressed patients’ perceptions of facial emotions in depressed and remitted states are associated with relapse: a longitudinal study. J Nerv Ment Dis. 1999;87:595-602.
22. Young AW, Rowland D, Calder AJ, et al. Facial expression megamix: tests of dimensional and category accounts of emotion recognition. Cognition. 1997;63: 271-313.
23. Harmer CJ, O’Sullivan U, Favaron E, et al. Effect of acute antidepressant administration on negative affective bias in depressed patients. Am J Psychiatry. 2009;166:1178-1184.
24. Harmer CJ, Shelley NC, Cowen PJ, Goodwin GM. Increased positive versus negative affective perception and memory in healthy volunteers following selective serotonin and norepinephrine reuptake inhibition. Am J Psychiatry. 2004; 161:1256-1263.
25. Harmer CJ, de Bodinat C, Dawson GR, et al. Agomelatine facilitates positive versus negative affective processing in healthy volunteer models. J Psychopharmacol. 2011;25:1159-1167.
26. Lemoult J, Yoon KL, Joormann J. Affective priming in major depressive disorder. Front Integr Neurosci. 2012;6:76.
27. Rinck M, Becker ES. A comparison of attentional biases and memory biases in women with social phobia and major depression. J Abnorm Psychol. 2005;114: 62-74.
28. Bradley B, Mathews A. Negative self-schemata in clinical depression. Br J Clin Psychol. 1983;22:173-181.
29. Chan SW, Goodwin GM, Harmer CJ. Highly neurotic never-depressed students have negative biases in information processing. Psychol Med. 2007;37:1281- 1291.
30. Bradley BP, Mogg K, Williams R. Implicit and explicit memory for emotion-congruent information in clinical depression and anxiety. Behav Res Ther. 1995; 33:755-770.
31. Arnone D, Horder J, Cowen PJ, Harmer CJ. Early effects of mirtazapine on emotional processing. Psychopharmacology (Berl). 2009;203:685-691.
32. Harmer CJ, Hill SA, Taylor MJ, Cowen PJ, Goodwin GM. 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.
33. Harmer CJ, Heinzen J, O’Sullivan U, Ayres RA, Cowen PJ. Dissociable effectsof acute antidepressant drug administration on subjective and emotional processing measures in healthy volunteers. Psychopharmacology (Berl). 2008;199: 495-502.
34. Grillon C, Duncko R, Covington MF, Kopperman L, Kling MA. Acute stress potentiates anxiety in humans. Biol Psychiatry. 2007;62:1183-1186.
35. Arnone D, McKie S, Elliott R, et al. Increased amygdala responses to sad but not fearful faces in major depression: relation to mood state and pharmacological treatment. Am J Psychiatry. 2012;169:841-850.
36. Godlewska BR, Norbury R, Selvaraj S, Cowen PJ, Harmer CJ. Short-term SSRI treatment normalises amygdala hyperactivity in depressed patients. Psychol Med. 2012;42:2609-2617.
37. Fu CH, Williams SC, Cleare AJ, et al. Attenuation of the neural response to sad faces in major depression by antidepressant treatment: a prospective, event-related functional magnetic resonance imaging study. Arch Gen Psychiatry. 2004; 61:877-889.
38. Phan KL, Coccaro EF, Angstadt M, et al. Corticolimbic brain reactivity to social signals of threat before and after sertraline treatment in generalized social phobia. Biol Psychiatry. 2013;73(4):329-336.
39. Kim MJ, Loucks RA, Palmer AL, et al. The structural and functional connectivity of the amygdala: from normal emotion to pathological anxiety. Behav Brain Res. 2011;223:403-410.
40. Murphy SE, Mackay CE. Using MRI to measure drug action: caveats and new directions. J Psychopharmacol. 2011;25:1168-1174.
41. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry. 1960; 23:56-62.
42. Montgomery SA, Asberg M. A new depression scale designed to be sensitive to change. Br J Psychiatry. 1979;134:382-389.
43. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry. 1961;4:561-571.
44. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16:606-613.
45. Watson D, Clark LA, Tellegen A. Development and validation of brief measures of positive and negative affect: The PANAS scales. J Pers Soc Psychol. 1988; 54:1063-1070.
46. von Zerssen D, Strian F, Schwarz D. Evaluation of depressive states, especially in longitudinal studies. In: Pichot P, ed. Psychological Measurements in Psychopharmacology. Basel, Switzerland: Karger;1974:189-202.
47. Spielberger CD, Gorusch RL, Lushene R, Vagg PR, Jacobs GA. Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychology Press; 1983.
48. Bond A, Lader M. The use of analogue scales in rating subjective feelings. Br J Psychol. 1974;47:211-218.


Keywords: anxiety; cognitive neuropsychology; depression; emotion; emotional processing