DOPING IN SPORT Introduction 1. During the Sydney Olympic Games in 2000, a Panel of the Court of Arbitration & Sport ("CAS") issued a decision in the case of the young Roumanian gymnast, Ms Raducan. That decision included the following: "The Panel is aware of the impact its decision will have on a fine, young, elite athlete. It finds, in balancing the interests of Ms Raducan with the commitment of the Olympic Movement to drugs-free sport, the Anti-Doping Code must be enforced without compromise."
Pone.0021881 1.8Altered Negative Unconscious Processing in MajorDepressive Disorder: An Exploratory NeuropsychologicalStudy Zhi Yang1, Jinping Zhao2, Yi Jiang3*, Chunbo Li4*, Jijun Wang4, Xuchu Weng3,5, Georg Northoff5,6 1 Key Laboratory of Behavioral Science, Institute of Psychology, Chinese Academy of Sciences, Beijing, China, 2 School of Life Science and Technology, University of Electronic Science and Technology of China, Chengdu, China, 3 Key Laboratory of Mental Health, Institute of Psychology, Chinese Academy of Sciences, Beijing, China, 4 Shanghai Mental Health Center, Shanghai Jiao Tong University School of Medicine, Shanghai, China, 5 Center for Human Brain Research, Hangzhou Normal University, Hangzhou, China, 6 Institute of Mental Health Research, University of Ottawa, Ottawa, Canada Objective: Major depressive disorder (MDD) has been characterized by abnormalities in emotional processing. However,what remains unclear is whether MDD also shows deficits in the unconscious processing of either positive or negativeemotions. We conducted a psychological study in healthy and MDD subjects to investigate unconscious emotionprocessing and its valence-specific alterations in MDD patients.
Methods: We combined a well established paradigm for unconscious visual processing, the continuous flash suppression,with positive and negative emotional valences to detect the attentional preference evoked by the invisible emotional facialexpressions.
Results: Healthy subjects showed an attentional bias for negative emotions in the unconscious condition while this valencebias remained absent in MDD patients. In contrast, this attentional bias diminished in the conscious condition for bothhealthy subjects and MDD.
Conclusion: Our findings demonstrate for the first time valence-specific deficits specifically in the unconscious processing ofemotions in MDD; this may have major implications for subsequent neurobiological investigations as well as for clinicaldiagnosis and therapy.
Citation: Yang Z, Zhao J, Jiang Y, Li C, Wang J, et al. (2011) Altered Negative Unconscious Processing in Major Depressive Disorder: An ExploratoryNeuropsychological Study. PLoS ONE 6(7): e21881. doi:10.1371/journal.pone.0021881 Editor: Manos Tsakiris, Royal Holloway, University of London, United Kingdom Received March 10, 2011; Accepted June 11, 2011; Published July 6, 2011 Copyright: ß 2011 Yang et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Support for this work was provided by the National Basic Research Program of China (973 Program, 2007CB512306 to XW), NSFC (30900366 to ZY;31070903 to YJ), The Science Foundation from Institute of Psychology, CAS (O9CX012001 to ZY), and CIHR, EJLB-CIHR, Michael Smith, and HDRF-ISAN to GN. Thefunders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: email@example.com (CL); firstname.lastname@example.org (YJ) negative emotions (see  for a review). In contrast, the findings inunconscious emotion processing are rather inconsistent. Some Emotion processing operates on a conscious level as well as in an studies reported deficits in unconscious emotional processing in unconscious (e.g., implicit and automatic) mode, with both being MDD [9–11] whereas others failed to find any changes in MDD .
associated with different neurobiological pathways [1–3]. A large For example, Mogg et al.  applied a Stroop task with emotional body of literature has focused on the conscious aspect of emotion stimuli that were backward masked and thus unconsciously processing as for instance in studies on emotional-cognitive regulation presented; MDD patients did not show any abnormalities in this and its abnormalities (e.g., [4–5]). In contrast, the unconscious aspect task (see also [13–14]). Similarly, several studies could not observe has been considered as the perception and earlier processing of the an attention bias with masked emotional stimuli in depressed emotion that precedes their cognitive regulation . What remains participants either [15–16]. A neuroimaging study  did not find unclear though is whether especially these different steps in emotion any difference in right amygdala in response to backward-masked processing, i.e., early (unconscious) and late (conscious), are valence- emotional faces between MDD patients and control subjects. In specific and thus different for positive and negative emotions in contrast, more recent analogous imaging studies did report deficits healthy subjects (for reviews, see ).
in the amygdala activity during masked faces [9–11]. Taken The situation is even more complicated in patients with major together, recent findings on unconscious emotional processing are depressive disorder (MDD) who suffer from an abnormal imbalance inconsistent with regard to the unconscious deficits in MDD patients between positive and negative emotions (see  for a recent review).
and it also remains unclear whether they are valence-specific.
MDD can be characterized by deficits in conscious emotion Due to the fact that the commonly applied technique of processing as it is, for instance, required in regulation of especially backward masking yielded rather mixed results in unconscious PLoS ONE www.plosone.org July 2011 Volume 6 Issue 7 e21881 Altered Negative Unconscious Processing in MDD emotional processing in MDD, we here adopted a different a mirror stereoscope mounted on a chinrest, each eye of the approach, continuous flashing interocular suppression (CFS) [18– participant could only see the frame on the same side, and the 20]. The CFS has been successfully and reliably used to tap into mirror stereoscope was adjusted so that the two frames were the implicit and automatic processing in the visual domain [21– comfortably fused together for the participant. The viewing 22]. Compared to the backward-masking technique, the CFS can distance was 40 cm (see  for setup of the equipment).
present information unconsciously throughout a relatively long The general experimental paradigm is shown in Figure 1. There viewing period (sometimes longer than 3 mins), potentially were two separate sessions for invisible and visible stimuli, each allowing for more robust and reliable unconscious processing containing 80 trials. At the beginning of each trial, a fixation cross . In order to investigate the unconscious processing in the (visual angle 0.8u60.8u) was presented to each eye for one second.
emotional domain, we combined the CFS with emotional faces of The following stimuli lasted for 800 ms in Experiments 1 and 3, positive and negative valences. Using this paradigm, we conducted and 200 ms in Experiment 2. The only difference between visible two experiments in healthy subjects using different stimulus and invisible sessions was the content of the stimuli: In the invisible durations (200 ms, 800 ms) of the viewing period (Experiments 1 session, a pair of faces with different emotional expressions was and 2). This served as the basis for a third experiment (Experiment presented to the participant's non-dominant eye (tested before 3) where we directly compared healthy and MDD subjects in the experiment), while a pair of identical high contrast dynamic noise emotional CFS to explore the valence-specific deficits in patches was presented to the dominant eye, so that the participant unconscious emotional processing in MDD.
only perceived the identical noise patches due to the stronginterocular suppression. For the visible session, a pair of faces with different emotional expressions was presented to both eyes and theparticipants could perceive the faces.
Each of the above face pairs was formed by grey-scale images of All participants were recruited at Shanghai Mental Health a positive and a negative emotional faces (4u66u of visual angle), Center. After a complete description of the study, written informed which were selected from happiness and sadness categories from consent was obtained from each participant. The protocol of this three actors/actresses in Ekman and Friesen's  pictures of study was approved by the Institute's Ethical Committees of both facial affect. The faces were masked using an ellipse so that the Shanghai Mental Health Center and Institute of Psychology, hair and background in the face images were excluded. The noise Chinese Academy of Sciences.
patches were masked in the same way to ensure their shape is the For Experiments 1 and 2, twenty healthy participants (13 same as the face images. The distance between the centers of the females, 24.064.2 years old) were recruited according to following two faces was around 5u. In half of the trials, the positive criteria: 13-term version of Beck Depression Inventory (BDI-13) emotional faces were presented to the left, and the negative ones to #4; Self-rating Anxiety Scale (SAS)#40; 17-term version of the right side. For the other half trials, the positions were reversed.
Hamilton Rating Scale for Depression (HAMD-17)#7; 14-term The above stimuli were followed by a 100-ms fixation cross, and version of Hamilton Rating Scale for Anxiety (HAMA-14)#7; then identical small Gabor patches (2.5u62.5u) were presented to normal or corrected-to-normal vision (tested with international both eyes for 100 ms. The positions of the Gabor patches were the standard visual testing chart); no history of psychiatric or centers of either the left or right previously presented faces. The neurologic (or medial) disease; and no substance abuse. Both Gabor patches were tilted one degree clockwise or counterclock- HAMD-17 and HAMA-14 have Chinese versions with good wise (randomized), and the participant was instructed to press one reliability and validity .
of the two buttons (2-alternative force choice) to indicate their For Experiment 3, twenty-three inpatients with MDD diag- perception of the orientation of the Gabor patches, regardless of nosed according to DSM-IV by two professional psychiatrists were which side the Gabor patches were presented to. The fixation recruited (13 females, 31.869.8 years old). Only subjects with would not end until the participants made their choice. Among all BDI-13 score.7 and HAMD-17.7 were included. All the 80 trials in each session, the presentation sides of the emotional participants were cooperative during the test without severe faces and that of the Gabor patches were balanced so that there suicidal ideation. The average HAMD score was 19.6 (SD = 9.3).
were 20 trials for each combination. The button-press and MDD patients with severe anxious symptoms (Self-rating Anxiety response time (RT) relative to the onset of the Gabor patch Scale (SAS).60, 14-item version of Hamilton Rating Scale for presentations was recorded.
Anxiety (HAMA-14).29) and/or other neurologic, psychiatric or Before the experiment, the participants were familiarized with medical disorders were excluded from the study. MDD patients the paradigm through a 50-trial training session. To ensure the were either medicated or not (serotoninergic drugs; see Table 1 for participants were never explicitly aware of the invisible faces, they were instructed to press a different key to reject the trial if they In addition, twenty-three healthy control participants (HC, 13 detected grey-scale images during the invisible session. If more females, 29.8610.1 years old) matched for age and gender with than two trials were rejected in the invisible session, the data for the MDD patients were recruited according to the criteria in the corresponding participant were excluded from further Experiment 1; this sample of healthy subjects was different from the one reported in Experiments 1 and 2. Demographic and Experiments 1 and 2 were exactly the same except that the clinical data for the MDD (further separated into medicated and duration of the face presentation was 800 ms in Experiment 1 and un-medicated) and the HC groups were listed in Table 1 (Part of 200 ms in Experiment 2. Experiment 3 was the same as the information was obtained after data screening).
experiment 1 but was applied to different participant populations,i.e., a new group of HC and a MDD group. For each session, the Stimuli and Procedure participants' responses were divided into two categories according Stimuli were programmed with the psychophysical toolbox [25– to the emotional valence of the faces that were on the same side 26] on MATLAB (The MathWorks, Natick, MA), and were with the Gabor patches. The stimuli categories were defined as presented on a 17-inch flat-panel monitor. Two square frames positive and negative emotion respectively. For each category, (10.7u610.7u) were displayed side by side on the screen. Through response accuracy (denoted as Accpos and Accneg) was calculated PLoS ONE www.plosone.org July 2011 Volume 6 Issue 7 e21881 Altered Negative Unconscious Processing in MDD Table 1. Clinical information summary of the participants.
Mean Illness Duration (Yrs) *Antidepressants (mg/d): Escitalopram(10); Sertraline(50); Paroxetine(20); Fluvoxamine(50); Fluoxetine(20); Venlafaxine(75); Citalopram(20); Mirtazapine(15); Doxepin(25); Trazodone (50); John's Wort Extracts (300).
doi:10.1371/journal.pone.0021881.t001 by dividing the number of correct responses by the total number of first performed to examine the gender and age difference between valid trials. The trials with RT longer than 1500 ms or shorter the two groups. After the within-group analyses (as used in than 100 ms were excluded. Within each session, the inter- Experiments 1 and 2), we further separated the MDD patients into category response accuracy difference, dAcc = Accpos2Accneg, was medicated and unmedicated groups, and conducted a repeated- used as a measure of attention preference. A positive or negative measure two-way ANOVA (visible/invisible6medicated/unmed- value indicated attention preference to the positive or negative icated) to examine the main effects and interactions between the emotional faces respectively.
two factors. Similarly, we conducted another repeated-measuretwo-way ANOVA (visible/invisible6HC/MDD) to examine the Statistical analysis difference between the healthy control and MDD participants.
In Experiments 1 and 2, the response accuracies for the positive and negative stimuli were separately tested against zero (using one- sample t-tests) for both the visible and the invisible conditions. Theattention preference, as measured by dAcc = Acc Experiment 1 (800 ms): Healthy subjects further tested (using paired t-tests) between the invisible and visible No participants were excluded according to the criteria conditions. In Experiment 3, before the above statistical analyses described in Methods. One-sample t-tests showed that the were conducted for each participant group, Chi-square tests were accuracy difference in the invisible session (dAccinvisible) was Figure 1. Schematic representation of the experimental paradigm for the invisible and visible conditions. In the invisible condition,dynamic noise patches were presented to the dominant eye and faces with happy and sad emotional expressions were presented to the other eye.
The duration of the face presentation was 800 ms for Experiments 1 and 3, and 200 ms for Experiment 2. After a 100-ms interval, participants wereinstructed to press one of two buttons as soon and accurate as possible to indicate the perceived orientation (clockwise or counter-clockwise) of aGabor patch presented for 100 ms. In the visible condition, the dynamic noise patches were replaced by the same pair of faces presented to the othereye.
doi:10.1371/journal.pone.0021881.g001 PLoS ONE www.plosone.org July 2011 Volume 6 Issue 7 e21881 Altered Negative Unconscious Processing in MDD significantly lower than zero [t(19) = 24.964, p = .000], indicating the difference between the invisible and visible sessions failed to observers' attentional preference for negative faces. This contrast- ed with the accuracy difference in the visible session (dAccvisible) p = .119] (See Figure 2, 200 ms group).
that was not significantly different from zero [t(19) = .178, p = .867]thus showing no attentional bias for positive or negative valences.
Experiment 3 (800 ms): Healthy and MDD subjects To further confirm the difference between visible and invisible Due to low response accuracy (lower than chance level), two sessions, we calculated the dAcc difference between both MDD patients had to be excluded. In the HC group, one conditions and the result was significant [dAccinvisible2dAccvisible, participant was excluded due to too few valid trials caused by t(19) = 22.595, p = .018] (see Figure 2). The difference is further abnormal response time. The age and gender ratio was matched demonstrated in the scatter plot of dAccinvisible and dAccvisible for between the two groups [age: t(41) = .524, p = .603, gender ratio: individual participants (see Figure 3A). In addition, we used a x2(1) = .020, p = .887]. The detailed demographic and clinical data standard bootstrapping procedure [29–30] to highlight the central are presented in Table 1.
tendency of dAccinvisible and dAccvisible. Specifically, from the A repeated-measure two-way ANOVA was performed to original participants, a bootstrapped sample with the same sample examine the difference between medicated and unmedicated size (i.e., 20 participants) was nonparametrically resampled with MDD subjects in the dAcc in both experimental conditions, visible replacement (i.e., a participant could be selected more than once).
and invisible; this yielded no significant statistics in either main The averaged dAccinvisible and averaged dAccvisible from this effects or interactions between medication groups and the bootstrapped sample was then plotted as a scatter plot (see experimental conditions [F(1, 19) = .17, p = .684 for main effect of Figure 3B). This procedure was repeated for 1000 times, and the experimental conditions, F(1, 19) = .325, p = .575 for experi- histograms representing population means and variations were mental condition6medication group interaction]. This indicated generated for the dAccinvisible and dAccvisible respectively. The the medication did not play an important role in the results, and histogram of dAccinvisible had a much higher kurtosis than that of therefore the two MDD groups were merged for further analyses.
dAccvisible while the mode of the dAccinvisible histogram was clearly The dAcc difference between the MDD and HC groups across lower than zero (See Figure 3B).
two experimental conditions (visible and invisible) was examinedby a repeated-measure two-way ANOVA. The results showed a Experiment 2 (200 ms): Healthy subjects significant interaction effect between the experimental conditions The same group of participants in Experiment 1 also participated and the subject groups [F(1, 41) = 4.537, p = .039].
in Experiment 2 which differed only in the presentation duration of Further simple effect analysis showed significant dAcc difference the stimulus (200 ms instead of 800 ms). This experiment was between MDD and HC groups in the invisible condition [F(1, undertaken to further examine whether the stimulus presentation 41) = 7.900, p = .008, see Figure 4]. The dAccinvisible for the MDD duration modulates the unconscious effects observed in the invisible group was not significantly different from zero [t(20) = 2.153, condition in Experiment 1. The dAcc p = .880], indicating that there was no bias between positive and invisible was not significantly different from zero [t(19) = 2.442, p = .663] while the dAcc negative valences in MDD patients in the invisible condition. In significantly higher than zero [t(19) = 2.519, p = .021]. Hence, the contrast, the dAccinvisible for the HC group was significantly shorter presentation time blurs the attentional preference to negative [t(21) = 24.296, p = .000], which was consistent with the negative emotions in the unconscious mode while it yields a positive findings in Experiment 1.
bias in the conscious mode (See Figure 2, 200 ms group). However, In the visible condition, however, there was no significant dAccvisible difference between HC and MDD groups [F(1,14) = .030, p = .874, see Figure 4]. Specifically, the dAccvisible forthe MDD groups was not significantly different from zero[t(20) = 24.296, p = .864], and so was the dAccvisible for the HCgroups [t(21) = .507, p = .618].
Similarly, we presented scatter plots for individual participants as in Experiment 1 to demonstrate that the difference between theMDD and HC groups in the invisible condition was more robustthan in the visible condition (see Figure 5A). To highlight the inter-group difference on dAccinvisible, bootstrap resampling proceduressimilar to that used in Experiment 1 were conducted within eachof the groups (see Figure 5B). The histograms representingpopulation means and variations showed significant differencebetween the two groups in the invisible condition but not in thevisible condition.
Figure 2. Comparison on attention preference in different This study investigated unconscious positive and negative sessions for healthy controls. The attention preference was indexedby difference in performance accuracy of the Gabor patch orientation emotion processing in both clinical MDD patients and healthy judgment task. A positive value indicates attention preference to happy controls. The main findings are: 1) Healthy subjects preferred to faces, and a negative value indicates attention preference to sad faces.
attend to positive emotional-valence stimuli in the visible, i.e., In the 800-ms session, the invisible condition revealed significantly conscious condition, while negative valences significantly attracted negative attention preference while the visible condition did not show more attention in the invisible, i.e., unconscious condition; and 2) significant valence preference. In the 200-ms session, no significant the unconscious attention preference to negative emotional valence preference was observed in the invisible condition; the visiblecondition showed a positive trend though.
valences remained absent in MDD patients which allowed to clearly distinguish them from the healthy group.
PLoS ONE www.plosone.org July 2011 Volume 6 Issue 7 e21881 Altered Negative Unconscious Processing in MDD Figure 3. Individual attention preference of healthy participants in the invisible and visible conditions. (A) Scatter plot of individualattention preference across the invisible and the visible conditions. Each dot represents the attention preference for a participant. The locations ofthe dots are determined by the visible (horizontal axis) and the invisible (vertical axis) conditions. The zero levels that indicate no attention preferenceare illustrated using dotted-lines. Except two participants, all others show a negative attention preference in the invisible condition, but no such trendappears in the visible conditions. (B) Scatter plot of bootstrapped sample means of the attention preference. 1000 datasets are resampled from theoriginal participants, each containing 20 participants, and the mean attention preference metrics for both invisible (horizontal axis) and visibleconditions (vertical axis) for each resampled dataset are represented by a point on the scatter plot. The locations of the points are separatelyprojected to the horizontal and the vertical axes, and the histograms are used to represent the distribution of the projected locations for the visibleand the invisible conditions respectively. In the invisible condition, the distribution is below the zero level and has a relatively small deviation,indicating a robust negative bias in population level, while the distribution for the visible condition is centered at the zero.
doi:10.1371/journal.pone.0021881.g003 Valence-specific unconscious emotional processing in  found that both positive and negative emotional reactions can be unconsciously evoked (also ); Jiang and He  showed Healthy subjects showed attention preference to sad facial that amygdala activity could be induced by fearful but not neutral expression in the unconscious condition (Experiment 1). This is in facial expressions during unconscious presentation. What remains accordance with various studies showing that emotional contents unclear though is whether this preference for emotions in the as distinguished from non-emotional ones are preferentially unconscious mode is valence-specific, meaning whether it pertains processed in unconscious condition. For example, Dimberg et al.
specifically for negative or positive emotions.
Our findings in healthy subjects demonstrated that processing in especially the unconscious mode is valence-specific by showing thespecific impact of negative emotions. However, such preference fornegative emotions was not observed in the conscious mode wherethe opposite valence, positive emotions, dominated. This raises thequestion for possible mechanisms of this preference for negativeemotions in the unconscious mode. One may assume that negativeemotions are processed with stronger and possibly also faster thanpositive emotions in the unconscious mode. This may also beneurobiologically plausible given that unconscious processing ofnegative emotions has been specifically associated with direct fastsubcortical pathways from the visual cortex to the amygdala (see[34,35] for a review). In contrast, conscious processing has beenrelated to relatively slower connections from the amygdala to theprefrontal cortex (for a review, see ) (see Figure 6a). However,this assumption remains speculative at this point awaiting furthersupporting evidence from future functional imaging studies.
In contrast to the preference for negative emotions in the unconscious mode, we observed a preference for positive emotions Figure 4. Comparisons of attention preference between in the conscious mode in Experiment 2. This is well in accordance healthy controls and MDD patients. In the invisible condition, with previous studies that also observed a similar positive attention the healthy controls showed significantly negative attention preference, bias in the conscious condition in healthy participants [36–39].
but the MDD patients did not. The difference between the two groups It should be noted though that we obtained such positive was significant. In the visible condition, the two groups behavedsimilarly.
preference only in our Experiment 2 whereas it was not observed in Experiment 1. We suspected that this was due to the 800-ms cue PLoS ONE www.plosone.org July 2011 Volume 6 Issue 7 e21881 Altered Negative Unconscious Processing in MDD Figure 5. Individual attention preference of healthy controls and MDD patients. (A) Scatter plot of attention preference of MDD patients(blue points) and healthy controls (red points) across the visible (horizontal axis) and the invisible (vertical axis) conditions. The horizontal and verticaldash lines represent the no-preference level for the visible and invisible conditions respectively. Most healthy controls showed negative attentionpreference in the invisible condition (most red points are below horizontal dash line), but there was no such trend in MDD patients (the blue pointsshow no obvious trend relative to the horizontal dash line). (B) Scatter plot of bootstrapped sample means of the two groups. For each participantgroup, 1000 datasets (with the same number of participants of the original group) are resampled from the original participants, and the meanattention preference metrics for both invisible and visible conditions for each dataset are represented by locations of the points on the scatter plot.
The locations are projected to horizontal and vertical axes, based on which the distributions of attention preference metrics for the visible andinvisible conditions are generated. The distribution for the MDD patients (blue histograms) and the healthy controls (red histograms) are clearlyseparated in the invisible condition but mixed together in the visible condition, showing a good separation between the two participant groups inthe invisible paradigm.
doi:10.1371/journal.pone.0021881.g005 duration as adopted and optimized for the unconscious condition [40–42]. Taken together, these findings support the proposed [18,27]. Due to this rather long duration, participants may have hypothesis that depressed individuals may not automatically (and directed their attention away from the stimuli in the conscious thus unconsciously) orient their attention towards negative condition thereby preventing the positive emotions form exerting information in the environment, but once such information has their full effect. In order to avoid this problem, we reduced the come to be the focus of their attention, they may have greater exposure duration to 200 ms in Experiment 2. As a result, the difficulty disengaging from it .
positive bias in the visible condition was observed as expected, Our results may also be clinically relevant in that unconscious which though impact the unconscious condition and its preference negative emotional processing may provide a novel and more for negative emotions in a negative way. Though seemingly a viable target for future psychotherapeutic and/or pharmacother- merely methodological problem, the impact of the duration time apeutic intervention than conscious emotion regulation strategies.
on the interaction with positive and negative emotions in both More specifically, it means that we have to target unconscious unconscious and conscious modes may point to an underlying processing rather than conscious processing as targeted in neurophysiological mechanism. More specifically, our results Cognitive Behavioral Therapy. Hence, our hypothesis, if con- suggest that the temporal duration of the stimulus may be central firmed in the future, may stipulate the development of more for inducing the neurophysiological mechanisms underlying the specific psychotherapeutic strategies.
valence-specific effects in both unconscious and conscious modes.
The same obviously holds for pharmacotherapeutic strategies.
This remains to be investigated in the future.
As discussed above, unconscious processing is mediated predom-inantly by fast processing in subcortical systems, while conscious Disturbed negative emotional unconscious processing in processing is rather related to slower processing in cortical regions.
depressed patients Interestingly, depressed patients indeed show major resting state In contrast to the healthy subjects, MDD patients did not show abnormalities in these subcortical regions ranging from the brain any preference towards negative emotions in the invisible stem (raphe nucleus, locus coerulus) to the dorsomedial thalamus, condition, i.e., the unconscious mode. Hence, their automatic the PAG, the tectum, the colliculi, and the amygdala (and and implicit negative emotional processing seems to be diminished ultimately to the cortex [6,9–11,17,43–45]).
in the unconscious mode. This complements and extends the How are the resting state abnormalities in these regions related many findings showing a negative bias in conscious emotion to the observed deficits in unconscious negative emotional processing in for instance emotion regulation and attention tasks processing? The abnormally high resting state activity  may PLoS ONE www.plosone.org July 2011 Volume 6 Issue 7 e21881 Altered Negative Unconscious Processing in MDD Figure 6. Conceptual schema combining the implications of the current findings into a general framework. As proposed by numericalstudies, the emotional stimuli are processed through three general phases, including analysis of stimulus features, recognition and response toemotion, and emotion regulation. The processing can be separated into conscious and unconscious parts. (A) Framework for healthy participants.
Studies on conscious emotion processing have suggested a positive preference in healthy subjects , we therefore mark the positive processing inconscious processing using a bold arrow and the negative processing using a narrow arrow. Our findings indicate that in the unconscious processingstage, the negative emotions may be processed through a stronger and faster pathway, as indicated by the bold arrow. (B) Framework for MDDpatients. Our findings also suggest that MDD patients may have a deficit in unconscious negative emotion processing (see the thinner arrow fornegative emotion in the unconscious part). This, in turn, may affect the conscious processing which then becomes shifted from positive to negativeemotion preference (see the bolder arrow for negative emotion in the conscious part) in order to compensate the hitherto incomplete processing ofnegative emotions in the deficient unconscious mode.
doi:10.1371/journal.pone.0021881.g006 affect first and predominantly the primary and fast unconscious puzzled about the relevance and significance of negative emotions; processing of emotional stimuli while it may exert less impact on this in turn may induce the often described emotion regulation the slower conscious processing. One would hence assume deficits and the abnormal attention bias towards negative emotions abnormally decreased rest-stimulus interaction  in subcortical in the conscious mode (see Figure 6b).
midline regions in MDD during especially negative emotional In conclusion, we here demonstrate for the first time negative processing. Such decreased rest-stimulus interaction may then be valence-specific effects in unconscious emotion processing in manifested in decreased preference for negative emotions in the healthy subjects and their disruption in MDD patients. This unconscious mode and that is exactly what we observed in our yields not only novel insights into unconscious processing in current study. Based on our findings, one may envision the general but suggests also that the often observed negative bias in developments of pharamcotherapeutic strategies (and other the conscious mode in MDD may result from a deficit in therapeutic tools like deep brain stimulation) that specifically unconscious negative processing.
target the subcortical fast tracks and their rest-stimulus interac-tion rather than the cortical slow connections. This however Author Contributions remains a scenario for the future.
Conceived and designed the experiments: ZY JZ YJ CL. Performed the How the deficits in unconscious negative processing are related experiments: ZY JZ CL JW. Analyzed the data: ZY JZ. Contributed to the conscious processing of emotions? Depressed patients show reagents/materials/analysis tools: XW. Wrote the paper: ZY YJ CL XW a deficit in unconscious emotion processing which may let them PLoS ONE www.plosone.org July 2011 Volume 6 Issue 7 e21881 Altered Negative Unconscious Processing in MDD 1. Adolphs R (2002) Neural systems for recognizing emotion. Curr Opin Neurobiol 24. Zhang M (1995) Handbook of Psychiatric Measures. Changsha: Hunan Science 12: 169–177.
and Technology Press.
2. LeDoux JE (1996) The emotional brain. New York: Simon and Schuster. 384 p.
25. Brainard DH (1997) The psychophysics toolbox. Spatial Vision 10: 433–436.
3. Phillips ML, Drevets WC, Rauch SL, Lane RD (2003) The neurobiology of 26. Pelli DG (1997) The VideoToolbox software for visual psychophysics: emotion perception I: Towards an understanding of the neural basis of normal Transforming numbers into movies. Spatial Vision 10: 437–442.
emotion perception. Biol Psychiatry 54: 504–514.
27. Yan X, Jiang Y, Wang J, Deng Y, He S, et al. (2009) Preconscious attentional 4. Joormann J, Gotlib IH (2010) Emotion regulation in depression: Relation to bias in cigarette smokers: A probe into awareness modulation on attentional bias.
cognitive inhibition. Cogn Emot 24: 281–298.
Addict Biol 14: 478–488.
5. Gotlib IH, Joormann J (2010) Cognition and depression: Current status and 28. Ekman P (1976) Pictures of facial affect. Palo Alto: Consulting Psychologists.
future directions. Annu Rev Clin Psychol 6: 285–312.
29. Efron B, Tibshirani R (1993) An introduction to the bootstrap. New York: 6. Tsuchiya N, Adolphs R (2007) Emotion and consciousness. Trends Cogn Sci 11: Chapman and Hall. 456 p.
30. Davison AC, Hinkley DV (1997) Bootstrap methods and their application.
7. Wiens S (2006) Subliminal emotion perception in brain imaging: Findings, Cambridge: Cambridge Univ Press. 592 p.
issues, and recommendations. Prog Brain Res 156: 105–121.
31. Dimberg U, Thunberg M, Elmehed K (2000) Unconscious facial reactions to 8. Northoff G, Hayes DJ (2011) Is our self nothing but reward? Biol Psychiatry, In emotional facial expressions. Psychol Sci 11: 86–89.
32. Ruys KI, Stapel DA (2008) The secret life of emotions. Psychol Sci 19: 385–391.
9. Dannlowski U, Ohrmann P, Bauer J, Kugel H, Arolt V, et al. (2007) Amygdala 33. Jiang Y, He S (2006) Cortical responses to invisible faces: Dissociating reactivity to masked negative faces is associated with automatic judgmental bias subsystems for facial-information processing. Curr Biol 16: 2023–2029.
in major depression: a 3 T fMRI study. J Psychiatry Neurosci 32: 423–429.
34. Phelps EA, LeDoux JE (2005) Contributions of the amygdala to emotion 10. Suslow T, Konrad C, Kugel H, Rumstadt D, Zwitzerlood P, et al. (2010) processing: From animal models to human behavior. Neuron 48: 175–187.
Automatic mood-congruent amygdala responses to masked facial expressions in 35. Leppanen JM (2006) Emotional information processing in mood disorders: A major depression. Biol Psychiatry 67: 155–160.
review of behavioral and neuroimaging findings. Curr Opin Psychiatr 19: 34–39.
11. Victor TA, Furey ML, Fromm SJ, Ohman A, Drevets WC (2010) Relationship 36. Hugdahl K, Iversen PM, Johnsen BH (1993) Laterality for facial expressions: between amygdala responses to masked faces and mood state and treatment in Does the sex of the subjects interact with the sex of the stimulus face? Cortex 29: major depressive disorder. Arch Gen Psychiatry 67: 1128–1138.
12. Mogg K, Bradley BP, Williams R, Mathews A (1993) Subliminal processing of 37. Kirita T, Endo M (1995) Happy face advantage in recognizing facial emotional information in anxiety and depression. J Abnorm Psychol 102: expressions. Acta Psychol 89: 149–163.
13. Lim S, Kim J (2005) Cognitive processing of emotional information in 38. McCabe SB, Gotlib IH, Martin RA (2000) Cognitive vulnerability for depression, panic, and somatoform disorder. J Abnorm Psychol 114: 50–61.
depression: Deployment of attention as a function of history of depression and 14. Yovel I, Mineka S (2005) Emotion-congruent attentional biases: The perspective current mood state. Cognitive Ther Res 24: 427–444.
of hierarchical models of emotional disorders. Pers Indiv Differ 38: 785–795.
39. Yoon KL, Hong SW, Joormann J, Kang P (2009) Perception of facial 15. Mogg K, Bradley BP, Williams R (1995) Attentional bias in anxiety and expressions of emotion during binocular rivalry. Emotion 9: 172–182.
depression: The role of awareness. Br J Clin Psychol 34(Pt 1): 17–36.
40. Gotlib IH, Krasnoperova E, Yue DN, Joormann J (2004) Attentional biases for 16. Mathews A, Ridgeway V, Williamson DA (1996) Evidence for attention to negative interpersonal stimuli in clinical depression. J Abnorm Psychol 113: threatening stimuli in depression. Behav Res Ther 34: 695–705.
17. Sheline YI, Barch DM, Donnelly JM, Ollinger JM, Snyder AZ, et al. (2001) 41. Gotlib IH, Kasch KL, Traill S, Joormann J, Arnow BA, et al. (2004) Coherence Increased amygdala response to masked emotional faces in depressed subjects and specificity of information-processing biases in depression and social phobia.
resolves with antidepressant treatment: an fMRI study. Biol Psychiatry 50: J Abnorm Psychol 113: 386–398.
42. Leppanen JM, Milders M, Bell JS, Terriere E, Hietanen JK (2004) Depression 18. Jiang Y, Costello P, Fang F, Huang M, He S (2006) A gender- and sexual biases the recognition of emotionally neutral faces. Psychiatry Res 128: 123–133.
orientation-dependent spatial attentional effect of invisible images. Proc Natl 43. Panksepp J (1998) Affective neuroscience: The foundations of human and animal Acad Sci U S A 103: 17048–17052.
emotions. New York: Oxford Univ Press. 480 p.
19. Fang F, He S (2005) Cortical responses to invisible objects in the human dorsal 44. Grimm S, Ernst J, Boesiger P, Schuepbach D, Hell D, et al. (2009) Increased and ventral pathways. Nat Neurosci 8: 1380–1385.
self-focus in major depressive disorder is related to neural abnormalities in 20. Tsuchiya N, Koch C (2005) Continuous flash suppression reduces negative subcortical-cortical midlines structures. Hum Brain Mapp 30: 2617–2627.
afterimages. Nat Neurosci 8: 1096–1101.
45. Alcaro A, Panksepp J, Witczak J, Hayes DJ, Northoff G (2010) Is subcortical– 21. Sterzer P, Kleinschmidt A, Rees G (2009) The neural bases of multistable cortical midline activity in depression mediated by glutamate and GABA? A perception. Trends Cogn Sci 13: 310–318.
cross-species translational approach. Neurosci Biobehav R 34: 592–605.
22. Brascamp JW, Knapen THJ, Kanai R, Van Ee R, Van Den Berg AV (2007) 46. Northoff G, Wiebking C, Feinberg T, Panksepp J (2011) The ‘resting-state Flash suppression and flash facilitation in binocular rivalry. J Vision 7: 1–12.
hypothesis' of major depressive disorder—A translational subcortical–cortical 23. Tsuchiya N, Koch C, Gilroy LA, Blake R (2006) Depth of interocular framework for a system disorder. Neurosci Biobehav R, In press.
suppression associated with continuous flash suppression, flash suppression, and 47. Northoff G, Qin P, Nakao T (2010) Rest-stimulus interaction in the brain: a binocular rivalry. J Vision 6: 1068–1078.
review. Trends in Neurosciences 33: 277–284.
PLoS ONE www.plosone.org July 2011 Volume 6 Issue 7 e21881
Impact of antiretroviral therapy on liver disease progression and mortality in patients co-infected with hiv and hepatitis c: systematic review and meta-analysis
Impact of antiretroviral therapy on liver disease progression and mortality in patients co-infected with HIV and hepatitis C Systematic review and meta-analysis Alexis Llewellyn, Mark Simmonds, Ginny Brunton, Social Science Research UnitUCL Institute of EducationUniversity College London EPPI-Centre report no. 2304October 2015 The authors are from the Centre for Reviews and Dissemination, University of York; and the EPPI-Centre, UCL Institute of Education, University College London.