Focus: How does one measure the early onset of antidepressant response?

Sakina RIZVI, HBSc
Department of Psychiatry
University Health Network
University of Toronto
Toronto, CANADA

How does one measure
the early onset of
antidepressant response?

by S. H. Kennedy, P. Giacobbe,
and S. Rizvi ,Canada

There is evidence to show that early symptomimprovement and early clinical response are predictive of favorable antidepressant outcome at the end of clinical trials. Early improvement describes a reduction in depression score from baseline of at least 20% after 2 weeks, and early response refers to a 50% reduction at or before 4 weeks. Typically, studies have relied on observer-rated scales such as the Hamilton Rating Scale for Depression or the Montgomery-Åsberg Depression Rating Scale. In a few instances, selfreport measures including the Beck Depression Inventory or the Hospital Anxiety and Depression Scale have been used for this purpose. However, these scales have generally not been validated for early use and repeated use over time frames of less than 7 days. Patient variables, including personality dimensions, may also bias frequent repeated measurements. Alternatively, the use of physiological measures, including neuroimaging, quantitative electroencephalography, and eye tracking shows promise in objectively detecting biological changes that precede mood improvement, and may distinguish between subsequent responders and nonresponders to treatment.With advances in communication technology, simple approaches such as daily diaries, text messaging, and interactive voice response systems can be employed to measure early “real-time” changes during treatment.

Medicographia. 2010;32:183-189 (see French abstract on page 189)

The historical opinion is that antidepressants take several weeks to be effective.1 For a number of decades, the possibility that some antidepressants may act more quickly than others has been a prominent research question.2,3 This debate extends beyond antidepressant medications, with reports that response to electroconvulsive therapy occurs more rapidly than response to pharmacotherapy,4 while cognitive behavioral therapy has a delayed trajectory of response compared with antidepressant medications.5

Aside from the nature of the treatment under investigation, patient variables may influence timing of response. These include age, sex, disease severity, personality, duration of episode and illness, as well as number of prior episodes.

There are also questions of methodology: how should “early onset” be defined? How frequently should patients be assessed? Which scales should be used, and what are the most appropriate statistical techniques to identify significant early advantages? These questions will form the focus of this review.

Defining early improvement and early response

There appears to be some consensus on a clinically meaningful definition of early symptom change. “Early improvement” has been operationalized as meaning a reduction of at least 20%fromthe baseline severity scoreoccurring within 2 weeks,6 usually based on the 17-item Hamilton Rating Scale for Depression (HAM-D17).7 Similarly, “early response” is considered to be at least a 50% decrease in the HAM-D17 score on or before week 4 of treatment, while responders after 4 weeks would be classified as “late responders.” The term “early stable responder” would apply to patients whomaintained at least a 50% reduction in symptom scores from 4 weeks to the end of a clinical trial.

These definitions have predictive validity in determining subsequent outcomes to antidepressant treatment. For example, Szegedi and colleagues8 completed a meta-analysis involving patients from over 40 clinical trials, who received either mirtazapine, a comparator, or placebo, and observed that 90% of all “stable responders” at the end of these trials came from the “early improvement” group, while only 11% of stable responders had not been in the early improvement group. These results support previous findings that improvement status at 2 weeks predicts ultimate response status.9,10

Symptom rating scales: clinician-administered

Early onset of action has primarily been assessed using HAMD17 (in some reports, using the 21-item version of HAM-D) or the Montgomery-Åsberg Depression Rating Scale (MADRS).11 These clinician-administered scales have been the gold standard in depression measurement for several decades and assess symptomseverity in the areas ofmood, anhedonia, sleep, anxiety, suicidality, libido, energy, and appetite over the preceding week.

There is conflicting evidence about the timing of individual symptom reduction across items. On the one hand, it is suggested that specific items of HAM-D have a different trajectory of improvement.12 In contrast, Stassen and colleagues13 reported that early score reductions occur for virtually all items of the scale, although certain items demonstrated greater percentage decreases than others (eg, suicidality, somatic anxiety, and energy levels).

It may be that trajectories of symptom improvement also vary across antidepressants. Findings suggest that early improvement (by 2 weeks) in psychic anxiety, psychomotor retardation, and suicidality, in particular, are predictive of remission with duloxetine, while changes in psychic and somatic anxiety are predictive of remission with escitalopram.14 In another report, treatment with desipramine resulted in reductions in psychomotor retardation and depressed mood after 1 week, and paroxetine treatment triggered early reductions in anxiety.15 Further replication is required to confirm these findings across antidepressants.

Several subscales of HAM-D have been validated and compared in depression trials. These include the Maier subscale,16 the Bech subscale,17 and the 7-item HAM-D (HAM-D7),18 although only the Maier subscale has been used to assess early onset of action. The Maier subscale comprises 6 items of the HAM-D: mood, guilt, work and activities, retardation, agitation, and psychic anxiety.16 In a post hoc analysis of a duloxetine and escitalopram trial, results demonstrated that failure to achieve a 20% improvement on the Maier subscale was highly predictive of unsuccessful treatment.14 Using the same measure as well as the HAM-D17, a 20% and 30% improvement with duloxetine treatment was observed at 21 days and 35 days, respectively, for both scales.19

The Clinical Global Impression (CGI) scale is another measure that has been used to evaluate timelines of early efficacy, based on a clinician’s overall judgment of severity and improvement.20 Although this simple measure does not assess specific symptoms, it has been found to correlate highly with depression scale scores.21 Response on this 7-point scale is often defined as a CGI-Improvement score of 1 or 2.22 Several trials have used this measure to assess early response, with similar results to those with HAM-D.23,24

Symptom rating scales: self-reported

Patient perspective is an important and often neglected issue in measuring symptom improvement. The scales most often used for this purpose are the Beck Depression Inventory-II (BDI-II),25 the Hospital Anxiety and Depression Scale (HADS),26 and the Quick Inventory of Depression Scale (QIDS).27 For the purpose of assessing early onset of action, clinician rating scales have primarily been employed. BDI-II, a 21-item scale, has been used in several pharmacotherapy trials to assess early improvement;28,29 however, it is the main measure utilized for tracking patient symptom progression during cognitive behavior therapy. Several studies have used BDI-II to show early improvement in this context.30-32 HADS and QIDS, 14-item and 16-item scales, respectively, have also not been utilized frequently in the literature to assess early response.33,34

_ Limitations of symptom scale measurement
While behavioral measures are a convenient way to assess early response, there are specific issues related to their utilization: whether to use a clinician-rated or a patient-rated scale, the frequency of measurement, as well as the time frame for measurement. Each of these variables may significantly influence the outcome of assessments.

Reports suggest that patient and clinician ratings may not necessarily exhibit high correlations.35,36 This demonstrates a discrepancy in the findings, whereby patients tend to rate themselves differently to their clinicians. Several studies suggest that the QIDS self report measure correlates highly with both its clinician-administered version as well as other clinician- rated scales such as HAM-D.37 However, personality dimensions and communication style can significantly weaken these correlations, as was demonstrated by Mattila-Evenden and colleagues38 using the Comprehensive Psychopathological Rating Scale. Severity of depression has also been shown to influence concordance between self-report and observerrated scores. Several lines of research suggest that there is less concordance between clinician and self-report ratings during an acute depressive episode compared with repeat tests after patient improvement,39-41 although it has been suggested that the increase in rating agreement after patients improve may be a statistical artifact.

The frequency of visits is also an important variable to consider, as an early event may be missed when visits are too infrequent, while too frequent visits may place unacceptable demands on most patients and result in the selection of a subgroup of patients who do not reflect the general population of depressed patients. There is also evidence to show that increased frequency of visits, particularly early on during the treatment, is associated with a higher rate of placebo response.42 This represents a complication in assessing early improvement, whereby more than one early time point of assessment could significantly reduce depression scores due to the therapeutic effect of visits.

The majority of scales discussed are validated on a set time frame (eg, the past week). However, issues arise when the frequency of assessment deviates from the time frame defined. HAM-D and MADRS were designed to assess the preceding 7 days; however, if a study necessitates more frequent visits within a week, this may compromise the validity of the scale in use. To date, there are no validation data on the use of HAM-D or MADRS more than once within 7 days. In cases where shorter time frames are required, the CGI-I may be a more appropriate measure, although specific symptom information will not be captured.

Physiological measurement of early improvement and response

Measurement scales such as HAM-D17 or MADRS are designed to gauge a wide variety of depressive symptomatology from diverse and distinct symptom clusters.43 As a result, the scores reflect a composite of multiple symptoms that improve at different rates.14 Methodologically, it is unclear whether early onset of antidepressant effect should be defined based on composite scores or on changes in individual clusters of symptoms. An alternative approach is to explore early changes in specific biomarkers (endophenotypes).

_ Emotional processing
Changes in the neurobiological substrates of emotional processing before and after antidepressant treatment may be a putative endophenotype for early response, and the measurement of these changes can be an indicator for early antidepressant response. There is consistent evidence that patients with depression exhibit biases in attending to, interpreting, and remembering negative emotional stimuli congruent with their mood state.44-46 In addition, there is evidence that antidepressant treatments are associated with acute changes in how people process emotional stimuli, and these effects precede any perceived benefit to mood. The facial expression recognition paradigm involves tasks that are able to tap into emotional processing, and features six basic emotions: happiness, surprise, sadness, fear, anger, and disgust, taken from individual characters in the Pictures of Facial Affect series.47 Several studies have used this task in healthy volunteers to show increased accuracy in identifying facial expressions of fear and happiness independent of reported mood state.48,49 Reductions in amygdala response to facial expressions of fear and increased activity in the fusiform gyrus to presentation of happy faces have also been observed, in some cases after just 7 days of treatment.50,51

Similar effects have also been reported in depressed patients using the same paradigm. Specifically, reports suggest that patients receiving an antidepressant demonstrate enhanced recognition of happy facial expressions, decreased reaction time to respond to positive versus negative self-referent items, as well as facilitated recall for positive items, and this in turn may be a predictor of antidepressant response at 6 weeks.52,53 In addition, preliminary evidence using the facial expression recognition paradigm in treatment-resistant depressed patients receiving vagus nerve stimulation, a neurosurgical procedure for depression, suggests that improvement in emotional processing may occur with the procedure prior to its clinical antidepressant effects.54 These results suggest that antidepressants as a class may share the ability to attenuate the cognitive biases seen in depression before changes are seen in the patient’s mood state, and that early detection of changes in emotional processing paradigms hold promise as a predictor of antidepressant response.

_ Quantitative electroencephalography
Quantitative electroencephalography (QEEG) is another potential method to assess early antidepressant response, as it has the ability to digitally measure electrical patterns (brainwaves) at the surface of the scalp reflecting cortical activity. After brainwaves are recorded, they are converted into num bers that are analyzed and transformed into a map of brain function. There is consistent evidence to suggest that changes in theta waves in the prefrontal cortex and anterior cingulate following antidepressant treatment may be associated with antidepressant response.55-57 A pooled analysis of QEEG recordings in patients with major depressive disorder (MDD) treated with fluoxetine, venlafaxine, or placebo found that responders at 8 weeks demonstrated changes in QEEG within 48 hours compared with those that did not respond to treatment.58 In fact, changes in QEEG measures within the first week of administration of an antidepressant medication have been reported to predict end-of-trial outcomes with an accuracy rate of over 70%.59

_ Selective attentional biases
Another promising indicator for detection of earlymood change is selective attention to emotionally-laden stimuli. Selective attention to negatively-valenced information supports and sustains the maladaptive patterns of information processing that are characteristic of depressive states,60,61 and recent evidence suggests that attentional biases causally alter emotional reactivity to stress.62 Recent papers63,64 have described and validated a new methodology to noninvasively measure changes in selective attentional biases in patients with MDD. In this method, depressed patients and healthy control subjects are asked to scan images with different thematic content while the pattern of their attentional deployment is continuously monitored by an infrared eye-tracking system mounted on the side of a computer monitor.63 Results from a report employing this task showed that subjects with MDD spent significantly more time looking at images with dysphoric themes compared with healthy controls, and that differences between the fixation times of the two groups was significantly correlated with the valence ratings. The authors concluded that subjects with MDD selectively attend to mood-congruent material, and that depression appears to influence the elaborative stages of processing when dysphoric images are viewed.

_ Limitations of physiological measurement
Promising methodologies to detect early changes in the processes that putatively contribute to an individual’s mood state include neuroimaging changes in regional activity following the presentation of an emotional facial expression or moodevoking words, changes in the preference to selectively look at images that are dysphoric versus pleasant, and early changes in measures of regional and global brain activity as measured by QEEG. All of these approaches represent noninvasive means to gauge early changes in brain functioning and the processing of emotional stimuli.

Specifically, there are a number of outstanding questions regarding the role of predicting antidepressant response from early changes in emotional processing. Firstly, the importance of the early changes in affective processing and behavior following acute antidepressant administration to the improvement in subjective and overt mood states is unknown. Antidepressants do not enhance mood in healthy individuals without a history of depression,65 therefore it is unknown whether the normalization of aberrant neural activity is necessary before clinical effects are seen, or is simply an epiphenomenon. This issue can be addressed through antidepressant trials, in which serial weekly assessments of depressive symptomatology are recorded concurrently with changes in emotional processing paradigms. This experimental design would allow the relationship of these changes to clinical outcome to be determined; namely, whether emotional processing changes invariably precede clinical improvement, what the time lag is between these changes and outcomes on scales, and whether those with a partial or no response to an antidepressant exhibit less change in these emotional processing paradigms.

Secondly, it remains unclear to what extent the immediate effects of antidepressants on the neural substrates of emotional processing interact with the known delayed effects of antidepressants on neurogenesis and neurotrophic factor expression. Studies are needed to assess whether the magnitude of the acute changes in emotional processing seen with antidepressants vary based on illness chronicity, and to what degree the changes relate to neuroimaging findings in depression, such as reduced hippocampal volume. Data to clarify all these issues would help address whether routine screening of these acute changes in emotional processing could be a promising screening paradigm for clinical outcomes.

The challenge for future studies with each of these methodologies is to elucidate their predictive value in a clinical setting. For example, do neuroimaging changes in response to sad facial expressions reliably precede clinically demonstrable changes in mood, and to what extent are residual biases still present in those who achieve antidepressant response? Are changes in selective attentional biases asmeasured by eye tracking patterns a state or trait marker of mood? Will novel QEEG markers be able to add to this methodology’s superior temporal resolution and achieve adequate spatial resolution to detect regional changes in brain activity that can predict antidepressant outcomes?

Future directions in measurement

The assessment of the onset of action of antidepressant treatments for MDD has generally been undertaken retrospectively during post-hoc analyses of clinical trials.66 Since clinical trials are adequately powered to detect a statistically significant effect at the end of the study, typically 4-8 weeks after the initiation of the treatment, standard antidepressant trials may lack adequate statistical power to declare any early changes as significant. A research agenda for future studies assessing the early onset of antidepressant treatments as its primary goal should be based on prospectively designed, controlled, randomized trials, which are adequately powered to detect early onset.66

Additionally, in order tomeasure early response,more frequent data capture is necessary; however, findings of increased placebo response with frequent visits pose a significant complication. Several ways around this issue concern increasing patient involvement in clinical trials. Lenderking and colleagues67 have been able to show that daily assessment via diary cards in patients receiving open-label fluoxetine had no effect on HAM-D or MADRS scores obtained in the clinic. In addition, daily diaries were able to detect therapeutic benefits earlier on than weekly assessments.

There is also encouraging evidence that up-to-date technology may aid in capturing more frequent data points. Currently in clinical trials, text message reminders to take the study drug are being implemented. In the same vein, brief questionnaires (eg, Maier, Bech subscale, or HAM-D7) regarding current depressive state could also be delivered via textmessage. Interactive Voice Response System technology has been in use for over a decade, which allows scales such as HAM-D to be automated over the phone.68 This is another relatively unexplored avenue for detecting early response that would enable the gathering of real-time data from patients.


Multiple lines of research have begun to challenge the firmly held clinical dictum that the effects of antidepressant treatments are delayed and that treatments require weeks before they exert their effects. A distinction needs to be made between delayed onset and delayed remission with antidepressants. The point has been made in the context of exploring the onset of action of antipsychotic medications, where although there is no debate that full therapeutic benefits take several weeks to realize, this by itself does not imply a delay in the onset of action.69

Without future investigations specifically designed to address the issue of time of onset of antidepressants, the time required to achieve improvement runs the risk of being misinterpreted as providing support for a delayed onset of action. Future investigations would benefit from the use of studies that are adequately powered to detect early clinical changes in outcome, refinements in the use of antidepressant measurement scales, and exploration of the relationship between the early neurobiological effects of existing antidepressants and clinical outcomes. _

1. Gelenberg AJ, Chesen CL. How fast are antidepressants? J Clin Psychiatry. 2000;61:712-721.
2. Taylor MJ, Freemantle N, Geddes JR, Bhagwagar Z. Early onset of selective serotonin reuptake inhibitor antidepressant action: systematic review and metaanalysis. Arch Gen Psychiatry. 2006;63:1217-1223.
3. Posternak MA, Zimmerman M. Is there a delay in the antidepressant effect? A meta-analysis. J Clin Psychiatry. 2005;66:148-158.
4. UK ECT Review Group. Efficacy and safety of electroconvulsive therapy in depressive disorders: a systematic review and meta-analysis. Lancet. 2003;361: 799-808.
5. Thase ME, Friedman ES, Biggs MM, et al. Cognitive therapy versus medication in augmentation and switch strategies as second-step treatments: a STAR*D report. Am J Psychiatry. 2007;164:739-752.
6. Stassen HH, Angst J, Delini-Stula A. Fluoxetine versus moclobemide: crosscomparison between the time courses of improvement. Pharmacopsychiatry. 1999;32:56-60.
7. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry. 1960;23:56-62.
8. Szegedi A, Jansen WT, van Willigenburg AP, van der Meulen E, Stassen HH, Thase ME. Early improvement in the first 2 weeks as a predictor of treatment outcome in patients with major depressive disorder: a meta-analysis including 6562 patients. J Clin Psychiatry. 2009;70:344-353.
9. Nierenberg AA, Farabaugh AH, Alpert JE, et al. Timing of onset of antidepressant response with fluoxetine treatment. Am J Psychiatry. 2000;157:1423-1428.
10. Stassen HH, Angst J, Hell D, Scharfetter C, Szegedi A. Is there a common resilience mechanism underlying antidepressant drug response? Evidence from 2848 patients. J Clin Psychiatry. 2007;68:1195-1205.
11. Montgomery SA, Asberg M. A new depression scale designed to be sensitive to change. Br J Psychiatry. 1979;132:382-389.
12. Farabaugh A, Mischoulon D, Fava M, et al. The relationship between early changes in the HAMD-17 anxiety/somatization factor items and treatment outcome among depressed outpatients. Int Clin Psychopharmacol. 2005;20:87-91.
13. Stassen H, Angst J, Delini-Stula A. Onset of improvement under fluoxetine and moclobemide. Eur Psychiatry. 1998;13:128-133.
14. Katz MM, Meyers AL, Prakash A, Gaynor PJ, Houston JP. Early symptom change prediction of remission in depression treatment. Psychopharmacol Bull. 2009;42:94-107.
15. Katz MM, Tekell JL, Bowden CL, et al. Onset and early behavioral effects of pharmacologically different antidepressants and placebo in depression. Neuropsychopharmacology. 2004;29:566-579.
16. Maier W, Philipp M, Gerken A. Dimensions of the Hamilton Depression Scale. Factor analysis studies. Eur Arch Psychiatry Neurol Sci. 1985;234:4170-4422.
17. Bech P, Gram LF, Dein E, Jacobsen O, Viitger J, Bolwig TG. Quantitative rating of depression states. Acta Psychiatr Scand. 1975;51:161-170.
18. McIntyre RS, Konarski JZ, Kennedy SH, Dickens SE, Bagby RM. Measuring depressive symptoms in the naturalistic primary-care setting. Int J Clin Pract. 2007;61:1278-1282.
19. Brannan SK, Mallinckrodt CH, Detke MJ, Watkin JG, Tollefson GD. Onset of action for duloxetine 60 mg once daily: double-blind, placebo-controlled studies. J Psychiatr Res. 2005;39:161-172.
20. Guy W. ECDEU Assessment Manual for Psychopharmacology. Rockville, MD: US Department of Health, Education, and Welfare; 1976.
21. Jiang Q, Ahmed S. An analysis of correlations among four outcome scales employed in clinical trials of patients with major depressive disorder. Ann Gen Psychiatry. 2009;8:4.
22. Quitkin FM, Rabkin JG, Ross D, et al. Identification of true drug response to antidepressants. Use of pattern analysis. Arch Gen Psychiatry. 1984;41:782-786.
23. Papakostas GI, Perlis RH, Scalia MJ, Petersen TJ, Fava M. A meta-analysis of early sustained response rates between antidepressants and placebo for the treatment of major depressive disorder. J Clin Psychopharmacol. 2006;26: 56-60.
24. Kennedy SH, Rizvi S. Agomelatine in the treatment of major depressive disorder: clinical effectiveness in perspective. CNS Drugs. In Press.
25. Beck AT, Steer RA, Garbin MG. Psychometric properties of the Beck Depression Inventory: twenty-five years later. Clin Psychol Rev. 1988;8:77-100.
26. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand. 1983;67:361-370.
27. Rush AJ, Trivedi MH, Ibrahim HM, et al. The 16-Item Quick Inventory of Depressive Symptomatology (QIDS), clinician rating (QIDS-C), and self-report (QIDS-SR): a psychometric evaluation in patients with chronic major depression. Biol Psychiatry. 2003;54:573-583.
28. Ambrosini PJ, Wagner KD, Biederman J, et al. Multicenter open-label sertraline study in adolescent outpatients with major depression. J Am Acad Child Adolesc Psychiatry. 1999;38:566-572.
29. Pollock BG, Perel JM, Nathan RS, Kupfer DJ. Acute antidepressant effect following pulse loading with intravenous and oral clomipramine. Arch Gen Psychiatry. 1989;46:29-35.
30. Brent DA, Holder D, Kolko D, et al. A clinical psychotherapy trial for adolescent depression comparing cognitive, family, and supportive therapy. Arch Gen Psychiatry. 1997;54:877-885.
31. Crits-Christoph P, Connolly MB, Gallop R, et al. Early improvement during manual- guided cognitive and dynamic psychotherapies predicts 16-week remission status. J Psychother Pract Res. 2001;10:145-154.
32. Hollon SD, DeRubeis RJ, Evans MD, et al. Cognitive therapy and pharmacotherapy for depression: singly and in combination. Arch Gen Psychiatry. 1992; 49:774-781.
33. Hunter MS, Ussher JM, Cariss M, Browne S, Jelley R, Katz M. Medical (fluoxetine) and psychological (cognitive-behavioural therapy) treatment for premenstrual dysphoric disorder: a study of treatment processes. J Psychosom Res. 2002;53:811-817.
34. Philip NS, Carpenter LL, Tyrka AR,Whiteley LB, Price LH. Varenicline augmentation in depressed smokers: an 8-week, open-label study. J Clin Psychiatry. 2009;70:1026-1031.
35. Fantino B, Moore N. The self-reported Montgomery-Asberg Depression Rating Scale is a useful evaluative tool in Major Depressive Disorder. BMC Psychiatry. 2009;9:26.
36. Uher R, Farmer A, Maier W, et al. Measuring depression: comparison and integration of three scales in the GENDEP study. Psychol Med. 2008;38:289- 300.
37. Rush AJ, Bernstein IH, Trivedi MH, et al. An evaluation of the quick inventory of depressive symptomatology and the Hamilton Rating Scale for Depression: a sequenced treatment alternatives to relieve depression trial report. Biol Psychiatry. 2006;59:493-501.
38. Mattila-Evenden M, Svanborg P, Gustavsson P, Asberg M. Determinants of self-rating and expert rating concordance in psychiatric out-patients, using the affective subscales of the CPRS. Acta Psychiatr Scand. 1996;94:386-396.
39. Fava GA, Kellner R, Lisansky J, Park S, Perini G, Zielezny M. Rating depression in normals and depressives: observer versus self-rating scales. J Affect Disord. 1986;11:29-33.
40. White J, White K, Razani J. Effects of endogenicity and severity on consistency and standard depression rating scales. J Clin Psychiatry. 1984;45:260-261.
41. Senra C, Polaino A. Concordance between clinical and self-report depression scales during the acute phase and after treatment. J Affect Disord. 1993;27: 13-19.
42. Posternak MA, Zimmerman M. Therapeutic effect of follow-up assessments on antidepressant and placebo response rates in antidepressant efficacy trials. Br J Psychiatry. 2007;190:287-292.
43. Shafer AB. Meta-analysis of the factor structures of four depression questionnaires: Beck, CES-D, Hamilton, and Zung. J Clin Psychol. 2006;62:123-146.
44. Surguladze S, Brammer MJ, Keedwell P, et al. A differential pattern of neural response toward sad versus happy facial expressions in major depressive disorder. Biol Psychiatry. 2005;57:201-209.
45. Fu CH, Williams SC, Brammer MJ, et al. Neural responses to happy facial expressions in major depression following antidepressant treatment. Am J Psychiatry. 2007;164:599-607.
46. Beck AT. The evolution of the cognitive model of depression and its neurobiological correlates. Am J Psychiatry. 2008;165:969-977.
47. Ekman P, Friesen WV. Pictures of Facial Affect. Palo Alto, Ca: Consulting Psychologists Press; 1976.
48. Harmer CJ, Bhagwagar Z, Perrett DI, Völlm BA, Cowen PJ, Goodwin GM. Acute SSRI administration affects the processing of social cues in healthy volunteers. Neuropsychopharmacology. 2003;28:148-152.
49. Harmer CJ, Mackay CE, Reid CB, Cowen PJ, Goodwin GM. Antidepressant drug treatment modifies the neural processing of nonconscious threat cues. Biol Psychiatry. 2006;59:816-820.
50. Murphy SE, Norbury R, O’Sullivan U, Cowen PJ, Harmer CJ. Effect of a single dose of citalopram on amygdala response to emotional faces. Br J Psychiatry. 2009;194:535-540.
51. Norbury R, Mackay CE, Cowen PJ, Goodwin GM, Harmer CJ. Short-term antidepressant treatment and facial processing: functional magnetic resonance imaging study. Br J Psychiatry. 2007;190:531-532.
52. 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.
53. Tranter R, Bell D, Gutting P, Harmer C, Healy D, Anderson IM. The effect of serotonergic and noradrenergic antidepressants on face emotion processing in depressed patients. J Affect Disord. 2009;118:87-93.
54. Critchley HD, Lewis PA, Orth M, et al. Vagus nerve stimulation for treatmentresistant depression: behavioral and neural effects on encoding negative material. Psychosom Med. 2007;69:17-22.
55. Bares M, Brunovsky M, Kopecek M, et al. Early reduction in prefrontal theta QEEG cordance value predicts response to venlafaxine treatment in patients with resistant depressive disorder. Eur Psychiatry. 2008;23:350-355.
56. Iosifescu D, Greenwald S, Devlin P, et al. Frontal EEG predictors of treatment outcome in major depressive disorder. Neuropsychopharmacology. 2009;19: 772-777.
57. Korb A, Hunter A, Cook I, Leuchter A. Rostral anterior cingulated cortex theta current density and response to antidepressants and placebo in major depression. Clin Neurophysiol. 2009;120:1313-1319.
58. Hunter AM, Muthén BO, Cook IA, Leuchter AF. Antidepressant response trajectories and quantitative electroencephalography (QEEG) biomarkers in major depressive disorder. J Psychiatr Res. In Press.
59. Hunter AM, Cook IA, Leuchter AF. The promise of the quantitative electroencephalogram as a predictor of antidepressant treatment outcomes in major depressive disorder. Psychiatr Clin North Am. 2007;30:105-124.
60. Dalgelish T, Watts EN. Biases of attention and memory in disorders of anxiety and depression. Clin Psychol Rev. 1990;10:589-604.
61. Mogg K, Bradley BP. A cognitive-motivational analysis of anxiety. Behav Res Ther. 1998;36:809-848.
62. MacLeod C, Rutherford E, Campbell L, Ebsworthy G, Holker L. Selective attention and emotional vulnerability: assessing the causal basis of their association through the experimental manipulation of attentional bias. J Abnorm Psychol. 2002;111:107-123.
63. Eizenman M, Yu LH, Grupp L, et al. A naturalistic visual scanning approach to assess selective attention in major depressive disorder. Psychiatr Res. 2003; 118:117-128.
64. Yu LH, Eizenman M. A new methodology for determining point-of-gaze in head-mounted eye-tracking systems. IEEE Trans Biomed Eng. 2004;51: 1765-1774.
65. Harmer CJ. Serotonin and emotional processing: does it help explain antidepressant drug action? Neuropharmacology. 2008;55:1023-1028.
66. Laska EM, Mallinckrodt CH, Mundt JC, et al. Assessing onset of treatment benefit in depression and anxiety: conceptual considerations. J Clin Psychiatry. 2009;70:1138-1145.
67. Lenderking WR, Hu M, Tennen H, Cappelleri JC, Petrie CD, Rush AJ. Daily process methodology for measuring earlier antidepressant response. Contemp Clin Trials. 2008;29:867-877.
68. Kobak KA, Greist JH, Jefferson JW, Katzelnick DJ, Mundt JC. New technologies to improve clinical trials. J Clin Psychopharmacol. 2001;21:255-256.
69. Agid O, Kapur S, Arenovich T, Zipursky RB. Delayed-onset hypothesis of antipsychotic action: a hypothesis tested and rejected. Arch Gen Psychiatry. 2003;60:1228-1235.

Keywords: early improvement; early response; major depressive disorder; Hamilton Rating Scale for Depression; quantitative electroencephalography; functional magnetic resonance imaging