EEG time–frequency analysis reveals blunted tendency to approach and increased processing of unpleasant stimuli in dysphoria

To date, affective and cognitive processing of emotional information in individuals with depressive symptoms have been examined through peripheral psychophysiological measures, event-related potentials, and time–frequency analysis of oscillatory activity. However, electrocortical correlates of emotional and cognitive processing of affective content in depression have not been fully understood. Time–frequency analysis of electroencephalographic activity allows disentangling the brain's parallel processing of information. The present study employed a time–frequency approach to simultaneously examine affective disposition and cognitive processing during the viewing of emotional stimuli in dysphoria. Time–frequency event-related changes were examined during the viewing of pleasant, neutral and unpleasant pictures in 24 individuals with dysphoria and 24 controls. Affective disposition was indexed by delta and alpha power, while theta power was employed as a correlate of cognitive elaboration of the stimuli. Cluster-based statistics revealed a centro-parietal reduction in delta power for pleasant stimuli in individuals with dysphoria relative to controls. Also, dysphoria was characterized by an early fronto-central increase in theta power for unpleasant stimuli relative to neutral and pleasant ones. Comparatively, controls were characterized by a late fronto-central and occipital reduction in theta power for unpleasant stimuli relative to neutral and pleasant. The present study granted novel insights on the interrelated facets of affective elaboration in dysphoria, mainly characterized by a hypoactivation of the approach-related motivational system and a sustained facilitated cognitive processing of unpleasant stimuli.

www.nature.com/scientificreports/ the passive viewing of pictures from the International Affective Picture System (IAPS) library 72 . The formulated hypothesis was twofold and was based on the abovementioned functional correlates of delta, theta and alpha bands. First, regarding affective disposition, the dysphoria group was expected to show a hypoactivation of the approach-related motivational system and, as suggested by the ECI model, a potential hypoactivation of the withdrawal-related motivational system. Specifically, the group with dysphoria was expected to show a smaller increase in delta band activity in response to pleasant and unpleasant vs. neutral pictures across spatially distributed cortical regions relative to controls. Also, considering that reviewed evidence supporting the role of alpha as a measure of the approach-related motivational system, the dysphoria group was expected to show a smaller alpha desynchronization in the left frontal and right parietal cortex in response to pleasant (but not neutral and unpleasant) stimuli relative to controls. Second, regarding cognitive processing, a facilitated cognitive processing of unpleasant and a reduced processing of pleasant stimuli was expected in the group with dysphoria. Namely, these processing patterns would be indexed by increased theta activity to unpleasant relative to neutral stimuli and to controls and by a reduced theta activity to pleasant relative to neutral pictures and relative to controls.

Methods
Participants. The present study was conducted within an extensive research project on vulnerability to depression and participants' EEG data have partially been described in a previous publication 33 . However, a distinct approach to data analysis was employed in the present study. A cohort of 85 Caucasian students at the University of Padua, Italy, voluntarily took part in the research project. The sample was medically healthy and free from psychotropic medication, as assessed with an ad-hoc anamnestic interview. In the present study, a group with dysphoria and a group without dysphoria were identified on specific criteria. Participants with dysphoria were identified by module A of the Structured Clinical Interview for DSM-5 (SCID 5-CV 73,74 ) assessing current and past depressive symptoms. Furthermore, the Beck Depression Inventory-II (BDI-II 75,76 ) was also employed for the assessment of depressive symptoms' severity. Based on the psychological assessment, 27 participants (5 males) who scored equal to or greater than 12 on the BDI-II and showed at least two present depressive symptoms, for at least two weeks, without meeting the diagnostic criteria for major depression, persistent depressive disorder, or bipolar disorder, were assigned to the group with dysphoria. Twenty-five participants (12 males) who scored equal to or lower than 8 on the BDI-II and had no history of depression or current depressive symptoms were assigned to the control group (i.e., without dysphoria) (see 33 , for more details on participants enrolled in the study). To ensure separation between groups with dysphoria and without dysphoria, participants who scored between 9 and 11 on the BDI-II were excluded from the present study (n = 17). Also, individuals without depressive symptoms but with at least one past depressive episode (i.e., remitted, see 77 ) were excluded from the present study (n = 16). With respect to demographic variables, the two groups included in the analyses (with dysphoria, without dysphoria) did not differ in terms of age (p = .645; dysphoria group: Mean (M) = 20. Participants were given 13 € for their participation. All participants read, understood, and signed informed consent. The research was conducted in compliance with the World Medical Association Declaration of Helsinki on research on human subjects and was approved by the Ethical Committee of Psychological Research, Area 17, University of Padua (prot. no. 3612).
Psychological measures. The Italian version of the mood episode module (module A) of the SCID-5-CV was employed as a reliable tool to assess the presence of dysphoria and to exclude individuals with major depression, persistent depressive disorder, or bipolar disorder. The SCID-5-CV was administered by a trained psychologist who had previous experience with administering structured clinical interviews. The Italian version of the BDI-II was also employed as a reliable measure of the severity of depressive symptoms in the past two weeks. The BDI-II is a self-report questionnaire composed of 21 items, each with a Likert scale of four-points and scores range from 0 to 63, where higher scores indicate greater depressive symptoms. In the Italian version, a score of 12 has been reported as the optimal cut-off score to discriminate between individuals with and without depressive symptoms 76 .
Experimental task and procedure. Twenty-four pleasant (e.g., erotic couples, sports), 24 neutral (e.g., household objects, neutral faces), and 24 unpleasant (e.g., attacking humans and animals) color pictures (600 × 800 pixels) were presented to participants. Highly arousing pleasant and unpleasant pictures were selected from the International Affective Picture System (IAPS; 72 ), since these have been observed to induce elevated psychophysiological changes (e.g. 7 ). Pleasant and unpleasant pictures were matched for normative arousal ratings which were significantly higher than for neutral pictures. The numbers of the selected IAPS pictures are listed in the supplementary material.
Pictures were shown for 6000 ms each in a semi-randomized sequence (i.e., no more than one stimulus in the same emotional condition had to be shown consecutively). The length of picture presentation was set to 6000 ms as the project included the registration of electrocardiographic activity to analyze cardiac deceleration (as reported in 33 ). Each picture was preceded by a 3000 ms interval where a white fixation cross was placed centrally on a grey screen. Participants were required to look at the central fixation cross and keep their gaze on the center of the screen. Picture presentation was followed by a variable intertrial interval (ITI) of 6000-8000 ms, during which a white fixation cross (identical to the 3-s baseline) was presented 33  www.nature.com/scientificreports/ Before the experimental session, participants were required to avoid alcohol consumption the day before and to avoid caffeine and nicotine the same day of the appointment. Upon arrival at the laboratory, after reading and signing written informed consent, participants were administered the ad-hoc anamnestic interview, the mood episode module (module A) of the SCID-5-CV, and the BDI-II. Then, participants were seated on a comfortable chair in a dimly lit, sound-attenuated room. After electrodes attachment and a 3-min resting-state period, six practice trials including two pleasant, two neutral, and two unpleasant pictures were provided. Then, participants underwent the emotional passive viewing task. The entire procedure took approximately 90 min. EEG recording. EEG data acquisition was accomplished using a computer running Eego software and using an Eego amplifier (ANT Neuro, Enschede, Netherlands). EEG was recorded using an elastic cap with 32 tin electrodes arranged according to the 10-20 System (Fp1, Fpz, Fp2, F7, F3, Fz, F4, F8, FC5, FC1, FC2, FC6, T7, C3, Cz, C4, T8, CP5, CP1, CP2, CP6, P7, P3, Pz, P4, P8, POz, O1, Oz, O2, and M1 and M2 [mastoids]), referenced online to CPz 33 . Both vertical and horizontal electrooculograms (EOGs) were recorded using a bipolar montage to monitor eye movements and eye-blinks. The electrode pairs were placed at the supra-and suborbit of the right eye and at the external canthi of the eyes, respectively 33 . Electrode impedance was kept below 10 kΩ. The EEG and EOG signals were amplified with Eego amplifier (ANT Neuro, Enschede, Netherlands), bandpass filtered (0.3-40 Hz), and digitized at 1000 Hz.
Data reduction and analysis. The EEG signal was downsampled to 500 Hz and re-referenced offline to a linked mastoids montage as implemented in EEGLAB 78 . Further processing was conducted in Brainstorm 79 . The EEG was filtered offline with a band-pass filter of 0.3-30 Hz and manually corrected for blink artifacts using independent component analysis (ICA). The EEG was then segmented into 6000 epochs, from 3000 ms before stimulus onset to 3000 ms after stimulus onset, to prevent boundary effects 17 . Each epoch was baselinecorrected by subtracting the mean pre-stimulus voltage between − 250 ms and − 50 ms. Segments that contained residual artifacts exceeding ± 70 μV (peak-to-peak) were excluded. By applying the a priori criteria of excluding individuals for whom more than 50% of trials were rejected, two participants (2 females) in the group with dysphoria were excluded due to excessive noise on the EEG recording and failed mastoid, respectively. Moreover, one participant in the group with dysphoria (1 female) and one in the group without dysphoria (1 male) were excluded due to excessive noise on electrode T7 and overall low-quality signal which precluded cluster-based time-frequency analysis. On the remaining sample, the artifact rejection led to an average ± SD acceptance of 19.0 ± 3.6 pleasant trials, 18.5 ± 3.0 neutral trials, and 19.1 ± 3.2 unpleasant trials in the dysphoria group, and of 18.7 ± 3.2 pleasant trials, 18.7 ± 3.2 neutral trials, and 19.0 ± 3.1 unpleasant trials in the control group. No statistically significant differences between groups or among emotional conditions in the average acceptance of pleasant, neutral, and unpleasant trials emerged (all ps > .15).
Time-frequency analysis was performed using Morlet wavelet transformation on individual trials for each 1-Hz frequency bin between 1 and 30 Hz, using a mother wavelet at 1 Hz with 3-s time resolution (full width at half maximum; FWHM). Time-frequency decompositions were then averaged for each participant and emotional condition, and the event-related spectral perturbation (ERSP) was computed as the change in power expressed in decibels (dB) relative to the baseline (− 900 to − 400 ms) in each frequency bin at each time point 17 . Then, data were grand averaged across each group for each emotional condition.

Statistical analysis.
A cluster-based approach has been conducted to control over type I error rate arising from multiple comparisons across electrodes and time points 80 . This approach is advantageous as it does not rely on assumptions about the distribution of the data or the theoretical underlying distribution of test statistics under the null hypothesis (i.e., Gaussian) 81 . Instead, the distribution is generated by the data itself, by iteratively shuffling the condition labels over trials (i.e., within-subjects) or over subjects (i.e., between-subjects) and recomputing the statistics. The shuffling is repeated thousands of times until a distribution of the test statistic value observed under the null hypothesis is generated. If the observed statistic value (i.e., the test statistic associated with the non-shuffled data) falls within the distribution of the null-hypothesis test statistic values, the null hypothesis cannot be rejected and this would indicate that the observed data could have been randomly generated 81,82 . Cluster-based correction assumes that a true effect should show a temporal and spatial extension, with neighbor sensors showing similar patterns 81 . With cluster-based correction, at each iteration of the nullhypothesis distribution generation, a threshold is applied to the time-frequency map, such that the outcome is units of clusters instead of single pixels (i.e., electrodes) 81 . In the present study, once thresholded values resulted from statistics across electrodes and time points were obtained, the differences within emotional conditions or between groups were shuffled pseudo-randomly 2000 times 17,33 . To obtain a 'null' distribution of effect sizes, the maximal cluster-level statistics (i.e., the sum of values across contiguously significant electrodes and time points at the threshold level) were extracted for each shuffle. For each significant cluster in the (non-shuffled) data, the cluster-corrected p-value was computed as the statistics of the proportion of clusters in the null distribution that exceeded the one obtained for the cluster in question 17,33 . The analysis was conducted with a − 100 to 1400 ms time window and clusters with a p corr < .05 were considered statistically significant.
To test within-group differences in event-related power changes among emotional categories (pleasant, neutral, unpleasant) and between-group (with dysphoria, without dysphoria) differences within each emotional category, cluster-based repeated measures ANOVAs and two-tailed unpaired t-tests were employed, respectively. The cluster-based statistical tests were run on the event-related delta (1-3 Hz), theta (4-7 Hz), and alpha (8-13 Hz) power over time-points in the − 100 to 1400 ms interval and a p < .05 criterion was employed to threshold the matrices 17  www.nature.com/scientificreports/ carried out to determine whether sex influenced the time-frequency activity in the clusters that emerged as significant between the two groups.

Results
Delta power. Differences Fig. 3. Specifically, the dysphoria group showed reduced delta in response to both pleasant and neutral stimuli relative to the group without dysphoria (see Fig. 3, panels b, c, e, and f). Unpaired t-test did not reveal any significant cluster for the difference between the groups within the unpleasant condition. Moreover, delta power within the significant clusters that emerged from the between-groups comparisons was not influenced by sex (neutral cluster, p = .125; pleasant cluster, p = .270).  Fig. 4b,c). Differently, the group without dysphoria revealed reduced theta power in response to unpleasant than neutral and pleasant stimuli (all ps ≤ .016; Fig. 5b,c).
Differences between groups in event-related theta power for each emotional category. Unpaired t-test conducted on event-related theta power did not reveal any significant cluster for the difference between the groups within each emotional condition (all ps ≥ .125).
Alpha power. Differences among emotional categories in event-related alpha power. The cluster-based analyses on event-related alpha power did not reveal any statistically significant cluster in testing possible withingroup differences (all ps ≥ .088).
Differences between groups in event-related alpha power for each emotional category. Unpaired t-test conducted on event-related alpha power did not reveal any significant cluster for the difference between the groups within each emotional condition (no cluster was detected; hence p-values were not generated). www.nature.com/scientificreports/

Discussion
The present study examined affective disposition and cognitive processing in dysphoria through the analysis of the time-frequency changes within delta, theta, and alpha bands during the exposure to emotional pictures 72 .
Regarding affective disposition, the dysphoria group was expected to show a hypoactivation of the approachrelated motivational system and, as suggested by the ECI model, a hypoactivation of the withdrawal-related motivational system. Second, the dysphoria group was expected to show selective facilitated top-down processing of unpleasant and a reduced processing of pleasant pictures. With respect to affective disposition, a pattern of increased event-related delta in response to all affective relative to neutral pictures emerged within both groups, indicating an affective modulation regardless of valence. Moreover, in line with the hypothesis, individuals with dysphoria showed an extended reduction of delta to pleasant pictures relative to the group without dysphoria. This finding possibly indicates a reduced emotional responding to pleasant images and a hypoactivation of the approach-related motivational system in individuals with dysphoria. Indeed, delta oscillations are linked to motivational processing, whereby an increase in its power indicates the identification of potentially rewarding cues (e.g. 42,46,49,62 ). Hence, reduced delta to pleasant images could denote reduced emotional responding to pleasant/rewarding stimuli in dysphoria. Overall, considering that reduced P3/LPP complex to pleasant images was observed in dysphoria 33 , two interrelated processes seem to www.nature.com/scientificreports/ characterize dysphoria: a hypoactivation of the approach-related motivation system and reduced motivated attention to positively valenced content. Additionally, the dysphoria group showed reduced delta to neutral pictures than the group without dysphoria, a pattern probably due to participants' motivational inertia, characteristic of depressive symptoms. Namely, it is plausible that the decreased motivation in dysphoria extended to non-relevant stimuli that did not elicit a saliency-detection process as prominent as in controls. Taken together, these findings provide support for the positive attenuation hypothesis in dysphoria. Conversely, the present findings are at odds with the negative potentiation hypothesis as well as with the reduced reactivity to unpleasant stimuli 20 . However, blunted reactivity to unpleasant stimuli might specifically be a manifestation of clinical depression. Furthermore, no significant difference between the dysphoria and the control group was found in the eventrelated alpha. This null finding may be due to different methodological approaches employed across studies. For instance, the present study differs in several methodological features from the few previous studies that employed a time-frequency approach 18,57 . Particularly, compared to a previous study 18 , here an even more rigorous statistical approach was employed, whereby group level analyses were conducted on distinct time windows identified through cluster-based analysis conducted within each group separately. Also, although reduced alpha desynchronization to pleasant stimuli was reported in depression, results are still inconsistent in the literature (for a review see 54 ).
Regarding affective cognitive processing, the two groups showed distinct patterns of theta power changes at the within-subjects level. Of note, these within-subjects differences occurred at distinct time windows, indicating potentially different processes occurring within the same stage of stimulus analysis. In the literature, two stages of emotional processing of theta power were identified: an early increase (~ 300 ms) related to automatic orienting and a later (after 300 ms) increase related to fine-grained top-down processing of salient stimuli 42,50 . Regarding the differences within the dysphoria group, an early increased in theta power for unpleasant pictures lasted until a later processing stage. On the other hand, within the control group, reduced theta for unpleasant pictures was evident only during a subsequent processing stage (836-1400 ms). It could be hypothesized that during the early stage of processing, individuals with dysphoria showed a preferential early orienting for unpleasant www.nature.com/scientificreports/ relative to both pleasant and neutral images. However, this effect was stable even after the early orienting stage, indicating that individuals with dysphoria performed a selective top-down processing towards unpleasant cues. Also, this pattern suggests that dysphoria may show a reduction in orienting towards pleasant pictures, which are processed as neutral ones. In contrast, controls showed a late reduction of top-down processing for unpleasant cues, suggestive for a conscious and adaptive regulation of these stimuli 83 . Consistently, a previous study on healthy participants reported that a late (1000-4000 ms) theta activity decrease was associated with reappraisal, a regulation strategy aimed at modifying the meaning of an emotional situation 83 . The late dampening of theta by reappraisal was interpreted as decreased prioritization of the stimuli by selective attention, following an initial evaluation of their affective saliency 83 . Hence, the within-subjects pattern in dysphoria not only is consistent with a facilitated processing of unpleasant cues, but it might indicate a lack of adaptive regulation strategies as compared to controls. Since theta band has been largely associated with high-order cognitive processes 84,85 , in future studies it would be interesting to investigate event-related theta while participants engage in complex affective cognitive task. From the current findings, the distinct role of delta and theta is supported. For instance, albeit speculative, not only they represent distinct functional correlates of affective processing, but they seem to be distinctively associated with the elaboration of pleasant and unpleasant content, respectively. In this regard, previous time-frequency studies on reward and loss processing have linked increased delta to reward sensitivity and increased theta to loss processing, describing them as two dissociable processes (e.g. 86,87 ). Interestingly, a previous study reported that depressive symptoms were prospectively predicted by diminished reward-related delta but not loss-related theta 87 . Despite these studies employed a different paradigm, the present findings support the perspective of a pleasantness-related delta band and unpleasantness-related theta band. The current findings on theta are novel and future studies are warranted to better disentangle its role in the top-down processing of affective stimuli in dysphoria. www.nature.com/scientificreports/ The time-frequency approach applied in the present study offers several methodological advantages compared to standard ERPs. Indeed, this approach allowed the separation of two peculiar measures of affective elaboration reflecting distinct processes occurring simultaneously, namely affective disposition and top-down processing. Furthermore, in addition to the analysis of evoked oscillations, time-frequency analysis also incorporates induced oscillations, known to carry important information about cognitive processes 40 .
Some limitations should be acknowledged. First, considering that the present study was based on a community sample and that dysphoria is more prevalent among the female population 3 , most of the participants within the dysphoria group belonged to the female sex. This sex unbalance might not allow the generalization of the findings to the male population and future studies are warranted to replicate the findings and increase their generalizability. Furthermore, although the emotional passive viewing task is a valid and widely used paradigm to study affective processing (e.g. 19,38 ), future studies that include specific experimental manipulations during the exposure to emotional stimuli are warranted to clarify the functional correlates of delta, theta and alpha frequency bands in affective tasks.
In conclusion, the present time-frequency study granted novel evidence on distinct but interrelated facets of affective elaboration in dysphoria, mainly characterized by a hypoactivation of the approach-related motivational system and a sustained facilitated cognitive processing along with reduced adaptive regulation of unpleasant stimuli. Considering that dysphoria is a condition known to considerably increase the risk of depression, these patterns of affective processing may represent quantitative measures allowing for early identification and treatment of depressed mood.

Data availability
The datasets analyzed during the current study are not publicly available due ethical concerns but are available from the corresponding author on reasonable request.