Assessing the Predictive Utility of Quantitative Electroencephalography Coherence in Adolescent Major Depressive Disorder: A Machine Learning Approach

Assessing the Predictive Utility of Quantitative Electroencephalography Coherence in Adolescent Major Depressive Disorder: A Machine Learning Approach | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Assessing the Predictive Utility of Quantitative Electroencephalography Coherence in Adolescent Major Depressive Disorder: A Machine Learning Approach Molly McVoy, Serhiy Chumachenko, Maia Gersten, Benjamin Wade, and 10 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6614439/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Improving early recognition and accurate diagnosis of Major Depressive Disorder (MDD) in childhood is a pressing concern. Quantitative electroencephalogram (qEEG) may be an effective, non-invasive diagnostic biomarker for MDD. Prior work by our team demonstrated decreased resting connectivity, as measured by qEEG coherence, in a heterogenous group of adolescents with MDD compared to age and gender matched health controls. This study explored qEEG coherence as a predictor of MDD diagnosis in a prospective, longitudinal sample of medication-free, adolescents with MDD vs healthy controls (HCs). Methods: Twenty-eight adolescents with MDD (Children’s Depression Rating Scale score ≥ 40), and twenty-seven age and gender matched HCs, (age 14-17, 78% female) received a baseline resting 32-channel EEG. Brain-wide coherence between channel pairs was calculated for the frequency bands (alpha, beta, theta, and delta) and compared between MDD youth and HC. Random forest classifiers were used to predict individual MDD status using baseline qEEG coherence. Models were trained and tested using 10-repeated, 10-fold cross validation and performance was evaluated with the area under the receiver operating characteristic curve (AUC-ROC). The contribution of individual predictors was assessed using permutation importance. Model significance was assessed using permutation testing (B=1000 resamples). Results: Random forest models predicted depression status with a trend-level of significance (mean AUC-ROC=0.65, p=0.08). Among the most predictive channel pairs, adolescent MDD was characterized by lower coherence in T7-P7 (p<0.05), Fz-Cz, and Fp2-F8 as well as higher coherence in P4-O2 and Cz-Pz. Conclusions: This study provides preliminary evidence that multivariate patterns of qEEG may inform diagnosis of adolescent MDD. Specific aberrant patterns of coherence within the default mode network and cognitive control network were characteristic of adolescent MDD. Ongoing work will seek to replicate these findings in a larger cohort. Psychiatry major depressive disorder adolescent qEEG quantitative electroencephologram biomarker connectivity Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Psychiatric disorders are one of the leading causes of morbidity and mortality in youth in the United States.(Bitsko et al., 2022 ; Lu and Keyes, 2023 ) Major depressive disorder (MDD) is the most common mood disorder in youth, affecting up to 20% of adolescents in the United States.(Mendelson and Tandon, 2016 ) Adolescent-onset MDD is associated with more severe and persistent symptoms than when diagnosed in adulthood(Harder et al., 2022 ; Johnson et al., 2018 ), higher levels of comorbidity and poorer psychosocial functioning.(Herzog et al., 2021 ; Strawn et al., 2020 ) The highly malignant effects of MDD place adolescents at risk for poor school performance, substance use disorders, self-injury, and in the worst cases, suicide.(Dwyer et al., 2020 ; Kupfer et al., 2012 ; Lu and Keyes, 2023 ; Walter et al., 2023 ) Suicide was the third leading cause of death in U.S. adolescents in 2021, taking the lives of 2,343 youth.(Centers for Disease Control and Prevention, 2021 ) To put this in perspective, in the last decade, more adolescents died from suicide than from cancer, cardiovascular disease, COVID-19 birth defects, lung disease, and influenza combined.(Centers for Disease Control and Prevention, 2021 ) Adolescence marks a period of re-organization and pruning of neuronal connections and is influenced by genes, environment, physical health and other unknown factors.(Gogtay et al., 2004 ; Griffin, 2017 ; Marsh et al., 2008 ; Whitford et al., 2007 ) MDD commonly arises during this time of rapid brain development, and may change the trajectory of this development.(Andersen and Teicher, 2008 ; Fuhrmann et al., 2015 ; Ismail et al., 2017 ; Yang and Tseng, 2022 ) While it is a unique time of risk, adolescence is also a time of unique promise and potential.(Fuhrmann et al., 2015 ; Ismail et al., 2017 ) Neuroscience research increasingly supports the notion that adolescence represents a “unique window of plasticity.”(Ismail et al., 2017 ) Consequently, if depression can be detected accurately and early, adolescence provides an opportunity to prevent or mitigate the devastating consequences of the disease.(Halfin, 2007 ; McGorry and Mei, 2018 ; Rice et al., 2017 ; Walter et al., 2023 ) Although MDD is common and potentially deadly, there are currently no objective biomarkers for this disease. Therefore, early recognition and accurate diagnosis are difficult, and rely on verbal reports from youth and parents.(McGorry et al., 2014 ; Walter et al., 2023 ) We report on a project that utilized quantitative electroencephalography (qEEG) as a diagnostic biomarker of adolescent MDD. Previous work by our team demonstrated decreased resting connectivity (qEEG coherence) in 25 MDD teens compared with 14 age and gender-match healthy controls.(McVoy et al., 2019a ) This decreased connectivity was most pronounced in the right prefrontal cortex, an area of the brain associated with the default mode network (DMN). The pilot study investigated a small number of adolescents, all on multiple psychiatric medications and with significant comorbidity, thus only allowing a limited analysis.(McVoy et al., 2019a ) We now present the baseline findings from a longitudinal study comparing those with MDD compared to age- and gender-matched healthy controls (HC). The study aims to assess whether qEEG and connectivity measures, specifically coherence, reveal differences between adolescents with MDD and their non-depressed, healthy peers. In addition, this study used machine learning (ML) methods to evaluate EEG markers of MDD. Previous studies have applied machine learning methods to identify diagnostic biomarkers in adults for depression using neuroimaging data, including functional magnetic resonance imaging (fMRI)(Yamashita et al., 2020 ) and EEG.(Liu et al., 2022 ; Mumtaz et al., 2018 ) To date, applications of ML to predict adolescent MDD have been developed using clinical history or electronic health records data; however, none have attempted to predict adolescent MDD using biological measures like qEEG. This study therefore presents a novel approach in the field. We hypothesized that the machine learning models will replicate the findings of our earlier research, and provide further evidence of the utility of qEEG as an intermediate biomarker of adolescent MDD. Methods Overall study description This study enrolled 55 adolescents aged 14–17, 28 with at least moderate MDD and 27 age- and gender-matched HC youth. MDD youth were followed over 16 weeks while HC youth had a one-time assessment visit. Clinical assessments included a diagnostic evaluation and standardized measures of depression, anhedonia, and anxiety. Following the baseline assessment, MDD youth were treated with evidence-based medication (fluoxetine or escitalopram) and a manualized brief therapy. The aim of the baseline study was to evaluate resting qEEG connectivity (coherence) in MDD cases vs. HCs. Subjects and recruitment: Subjects were recruited at University Hospitals Cleveland Medical Center (UHCMC), a tertiary care academic medical center, from both child psychiatry offices and general pediatrics’ practices. MDD adolescents aged 13–18 were included if they met DSM-5 criteria for Major Depressive Disorder (MDD) of moderate severity. Diagnosis and severity were determined through clinical evaluation by a Board Eligible/Board Certified Child Psychiatrist and the Children’s Depression Rating Scale- Revised (CDRS-R) ≥ 40. Subjects were scheduled to begin clinical treatment with escitalopram or fluoxetine for MDD. We excluded subjects with comorbid diagnoses that have reportedly shown differences on qEEG(Chan and Leung, 2006 ; Coben et al., 2008 ; di Michele et al., 2005 ; Fonseca et al., 2008 ; Kim et al., 2015 ; McVoy et al., 2019b ; Moeini et al., 2015 ), including attention deficit hyperactivity disorder (ADHD), bipolar spectrum disorder, autism spectrum disorder, primary psychotic disorder, active substance use disorder, post-traumatic stress disorder, and neurologic conditions including epilepsy, meningitis, history of brain surgery, shunts. In addition, treatment resistant youth, defined as having failed 2 or more trials of treatment with a selective serotonin reuptake inhibitor (SSRI) were excluded. Finally, the following medications were also exclusionary for MDD youth: benzodiazepines, atypical antipsychotics, anti-epileptics, psychostimulants, and any antidepressant medications. HC adolescents were age and gender matched to MDD youth, did not meet criteria for any DSM-5 psychiatric disorder (as assessed by a Board Eligible/Board Certified Child Psychiatrist), and did not have 1st degree relatives with a history of recurrent unipolar depression, mania, hypomania or psychosis or 2nd degree relatives with a history of mania, hypomania or psychosis. The same neurologic and medication exclusion criteria as applied to MDD adolescents was applied to HC adolescents. All subjects and their caregivers provided written consent. The study was approved by the local Institutional Review Board (IRB). Measures MDD and HC youth completed the follow assessment measures at screen. Children’s Depression Rating Scale-Revised (CDRS-R) : MDD severity was assessed with the CDRS-R.(Mayes et al., 2010 ; Poznanski and Mokros, 1996 ) The CDRS-R was originally derived from the Hamilton Depression Rating Scale, and is used for children aged 6–17.(Poznanski and Mokros, 1996 ) It is a 17-item scale, with items ranging from 1 to 5 or 1 to 7; total possible scores range from 17 to 113. A cut off score of ≤ 28 equates to minimal or no symptoms of depression, whereas a score of ≥ 40 indicates clinically significant depression. Snaith-Hamilton Pleasure Scale (SHAPS) : The SHAPS(Snaith, 1993 ) assesses anhedonia, or the inability to experience pleasure in normally-pleasant experiences. The SHAPS has 14 self-report Likert-scale items with the option to choose Strongly Disagree, Disagree, Agree, or Strongly Agree for each item. When scoring, each “Disagree” endorsement is given 1 point, and each “Agree” endorsement is given 0 points. Total scores range from 0–14, with higher scores indicating higher anhedonia. Pediatric Anxiety Rating Scale (PARS) : Anxiety was assessed with the PARS,(Riddle et al., 2002 ) a 50-item symptom checklist of items grouped into categories of: Social Interactions or Performance Situations (9 items), Separation (10 items), Generalized (8 items), Specific Phobia (4 items), Physical Signs and Symptoms (13 items), and Other (6 items). Symptoms are scored on 7 dimensions of severity, using a 6-point scale (0 for none, and 1–5 for minimal to extreme) for each dimension, and these are added to get a total score. Total scores range from 0 to 35, with higher scores indicating higher severity of anxiety. A score of ≥ 10 indicates clinically-significant anxiety, and scores ≥ 20 indicate severe anxiety.(Riddle et al., 2002 ) Clinical Global Impression-Severity (CGI-S:): CGI-S(Busner, 2007) uses clinical judgement to rate disease severity on a seven-point scale, as follows: 1 = normal, not at all ill; 2 = borderline mentally ill; 3 = mildly ill; 4 = moderately ill; 5 = markedly ill; 6 = severely ill; 7 = among the most extremely ill patients. Columbia Suicide Severity Scale (CSSRS) : Suicide risk was assessed with the CSSR-S(Posner et al., 2011 ), a screening tool that measures suicidality across 4 constructs: severity of ideation, intensity of ideation, behavior, and lethality. The severity subscale rates suicidal ideation on a 5-point scale: 1 = wish to be dead, 2 = nonspecific active suicidal thoughts, 3 = suicidal thoughts with methods, 4 = suicidal intent, and 5 = suicidal intent with plan. The intensity subscale rates frequency, duration, controllability, deterrents, and reason for ideation, each on a 5-point scale. The behavior scale rates actual, aborted, and interrupted attempts, as well as preparatory behavior and nonsuicidal self-injurious behavior on a nominal scale. The lethality subscale measures actual attempts on a 6-point scale. When actual lethality scores zero, potential lethality of attempts is rated on a 3-point scale. EEG data collection EEG data were acquired using a Discovery 24 EEG amplifier that has been approved by the U.S. Food and Drug Administration. Resting EEGs were recorded while subjects lay quietly with their eyes closed in a sound-attenuated room. Subjects were alerted frequently to avoid drowsiness and were instructed to remain still and inhibit blinks or eye movements during each recording period. EEGs were recorded using a 32-channel enhanced version of the International 10–20 System of Electrode Placement, using the montage in Fig. 1.(McVoy et al., 2019a ) qEEG analysis: We developed a semi-automatic tool to compute the power spectral density (PSD) and coherence measures for each subject. Further details of this tool and calculation of coherence are described in the pilot paper.(McVoy et al., 2019a ) The default segment length was set at 10 seconds for analysis, and the tool automatically analyzed the six eye channels (LOC-LSO, LOC-LIO, LSO-LIO, ROC-RSO, ROC-RIO, RSO-RIO) to detect segments with eye movements. PSD and coherence were then calculated for the selected EEG channels and for the delta (0.5–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), and beta (12–20 Hz) frequency bands. PSD and coherence were averaged across epochs for use in data analysis, as described below. Data Analysis: Random forest classifier (RF)(Breiman, 2001 ) models were used to predict individual MDD status (CDRS-R ≥ 40) using baseline EEG coherence. Previous studies have demonstrated that tree-based ensemble models like the RF model perform well on high-dimensional(Breiman, 2001 ) and tabular(Grinsztajn et al., 2022 ) style data structures compared to deep learning and other classical machine learning methods and are able to flexibly detect key interactions between predictors. We trained and tested RF models with 1000 underlying trees using 10-repeated, 10-fold cross validation. Our primary metric to evaluate model performance was the area under the receiver operating characteristic curve (AUC-ROC). Secondary evaluation metrics included the area under the precision-recall curve (AUC-PR), sensitivity, specificity, and positive and negative predictive values. Performance metrics were averaged across all 100 iterations of cross validation. The contribution of individual predictors was assessed using permutation importance.(Breiman, 2001 ) The significance of the model was assessed using permutation testing with B = 1000 resamples. Models were constructed in R version 4.3.0 using tidymodels.(Kuhn and Silge, 2022 ) Results Overall sample results Demographic and clinical variables are presented in Table 1 and a consort diagram of participants is presented in Fig. 2. Mean age of youth was 15.36 (SD = 1.16), N = 5855, 78% (N = 43) female. Of the MDD participants, 27 of 28 (96%) were experiencing a current major depressive episode, and 6 (21%) had a history of recurrent depressive episodes, as assessed by the MINI-KID. Of currently active diagnoses, 32% (N = 9) had panic disorder, 39% (N = 11) had agoraphobia, 43% (N = 12) had social anxiety, 18% (N = 5) had specific phobia, 25% had generalized anxiety disorder. Depression severity of the MDD participants, as measured by CDRS, was moderate, with an average of 63.1 (Range 40–84). Anxiety severity, as measured by PARS, was clinically significant with an average score of 15.5, with scores ranging from zero to 28. Six participants met the cut-off of ≥ 20 for severe anxiety. Suicidality, as measured by the MINI-KID, was present at baseline in 39% of MDD participants (N = 11), with 18% (N = 5) having a lifetime attempt, and 50% having no history of suicidality. Of the 11 with current suicidality, 5 were deemed low risk, 4 medium risk, and 2 high risk. See Table 1 for details of depression and suicidality measures within the MDD sample. Suicidal ideation, as measured by the CSSR-S, was present in 21 (75%) participants, with average ideation severity being 2.4 (Range 1–5) and average ideation intensity being 11.0 (Range 0–23). 14% of MDD (N = 4) participants had a previous suicide attempt, 1 of which had four previous attempts. 3%(N = 1) of MDD participants were endorsing suicidal behavior at the time of baseline assessment. 50% (N = 14) of participants had no previous trials of SSRI medication. The average number of previous trials among MDD participants was 0.73 (SD = 0.83), with a maximum observed trial of 3. Model Results Random forest models predicted depression status with a trend-level of performance (mean AUC-ROC = 0.65 [SD = 0.26], p = 0.08). Sensitivity (mean = 0.65 [SD = 0.28], p = 0.01) and negative predictive value (NPV; mean = 0.67 [SD = 0.24], p = 0.03) were both significant while specificity (mean = 0.58 [SD = 0.30], p > 0.1) was non-significant and positive predictive value (PPV; 0.63 [0.24], p = 0.07) was a trend-level of significance. Figure 3 illustrates the ROC curve for the model. See Table 2 for an outline of performance metrics. In descending order, the five most important features contributing to the prediction of MDD status were T7-P7 coherence beta, P4-O2 coherence beta, Cz-Pz coherence beta, Fz-Cz coherence delta, and Fp2-F8 coherence alpha. Figure 4 illustrates feature permutation importance for the top 10 most informative predictors. In general, lower T7-P7 coherence beta, Fz-Cz coherence delta, and Fp2-F8 coherence alpha were indicative of MDD. Conversely, higher P4-O2 coherence beta and Cz-Pz coherence beta were predictive of MDD. Figure 5 illustrates boxplots of the baseline coherence measures for the top 10 most important features. Post hoc tests were conducted to examine differences between MDD and HC youth within the predictive features. As the coherence data for both populations were not normally distributed, Wilcoxon rank sum tests were performed and revealed significant differences in features T7-P7 coherence beta (p < 0.001), Fz-Cz coherence delta (p < 0.05), and Cz-Pz coherence beta (p < 0.05). No significant differences were found in P4-O2 coherence beta (p = 0.27) or Fp2-F8 coherence alpha (p = 0.27). Following correction for multiple comparisons with the false discovery rate method (FDR), T7-P7 coherence beta remained significant (p < 0.01). Discussion We demonstrate a preliminary replication of our previous findings that measures of qEEG coherence differ in youth with MDD compared with HC in an expanded, comorbidity-free, treatment-naïve sample of MDD youth. Our RF classifier yielded trend-level significance in predicting adolescent MDD status using pretreatment qEEG measures. Our model had a significantly high sensitivity but non-significant specificity, suggesting that it performed better at identifying adolescents with MDD than HCs. The most informative qEEG measures in our model were biologically plausible channel pairs spanning regions often implicated in MDD including nodes of the default mode network (DMN) and cognitive control network (CCN)(Zhao et al., 2023 ). The DMN is primarily involved with emotional processing, self-referential thinking, rumination, as well as other aspects of cognition(Zhao et al., 2023 ). It is typically found to be hyperactive in MDD, though certain regions of the DMN may exhibit hypoactivity in connection with nodes of other functional networks.(Kaiser et al., 2015 ) The CCN is responsible for executive functions, such as working memory, attention, and planning(Breukelaar et al., 2017 ), and is also primarily hypoactive in depression.(Hack et al., 2023 ; Jiao et al., 2020 ) The most predictive and significant channel pair was T7-P7, which corresponds to the superior temporal gyrus (STG) and inferior lateral occipital cortex (LOC) respectively. The STG is associated with social and emotional processing(Lee et al., 2021 ; Takahashi et al., 2010 ), with previous research noting decreased activity in this region in response to sad stimuli in adult subjects with MDD.(Fitzgerald et al., 2008 ) It is also a primary component of the DMN. The LOC is a part of the visual association cortex that integrates visual information and has been associated with cognitive deficits or aberrant functioning in depressions.(Guan et al., 2021 ; Wu et al., 2023 ) Reduced activity between these two areas may reflect the diminished social and emotional processing observed in depression, particularly paying attention to and reacting to emotional visual stimuli like sad or threatening faces.(Fitzgerald et al., 2008 ) These deficits may be particularly relevant in adolescence, as social matters become increasingly salient. Prior to correction, two other channel pairs were found to be significant and shared an electrode, Fz-Cz and Cz-Pz. However, despite this common location, there were differences in the direction of the coherence comparisons between MDD and HC adolescents. While Fz-Cz coherence was decreased in MDD youth, Cz-Pz coherence was increased. Cz roughly corresponds to the precentral gyrus, while Fz and Pz may reflect the medial prefrontal cortex (mPFC) and precuneus respectively. Both the precuneus and mPFC are important nodes of the DMN, with the precuneus being linked to complex cognitive functions such as autobiographical memory(Utevsky et al., 2014 ) while the mPFC is associated with emotion regulation, social functioning, learning, stress response, and various cognitive functions.(Bittar and Labonté, 2021 ) The precentral gyrus is more associated with the CCN rather than the DMN and is typically involved in higher order cognitive functions, like selective attention and working memory.(Williams, 2016 ) When considering its link to each of the aforementioned regions, it is possible that decreased coherence in Fz-Cz could be related to the cognitive deficits observed in depression, such as decreased cognitive flexibility, emotion regulation, and concentration, while increased coherence in Cz-Pz may reflect the dysfunction of higher order cognitive processes, for instance self-referential thinking that may become pathological rumination. Future research should explore whether the precuneus and mPFC are negatively associated with each other, particularly when considering the precentral gyrus as a central node or seed. While P4-O2 was identified as the second most important feature in our model, there were no significant differences in the coherence values between MDD and HC youth. This is surprising, given that P4 corresponds to the inferior parietal lobule (IPL), an important node of both the DMN and the CCN. In summary, we identified decreased qEEG coherence in patients with MDD versus HCs in several dyads (T7-P7, Fz-CC), but increased coherence in Cz-Pz. Notably, one of the dyad locations was empirically replicated between the two studies (Cz-Pz) for MDD youth, hinting at potential consistent localization within the DMN that could strengthen a hypothesis of dysregulation in behavioral inhibition pathways leading to initial development of MDD.(Ho et al., 2015 ; Willinger et al., 2024 ) However, the direction of these comparisons differed between the two studies. In our pilot, we found, decreased coherence in Cz-Pz, but conversely increased coherence in this sample. In addition, our current sample identified lower T7-P7 coherence in MDD youth as compared to HC, a finding not observed in our pilot study. However, as stated previously these locations are consistent with the DMN(Das et al., 2022 ) and continue to support the theory that aberrant functioning of the DMN is characteristic of adolescent MDD. Clinically, qEEG continues to demonstrate feasibility and practicality as a potential biomarker in youth MDD and this research confirms differences are seen in resting connectivity between youth with MDD compared to HC, even in a medication free sample. Further work continues to be needed before incorporating qEEG as part of routine clinical care, including investigating qEEG markers of connectivity as it relates to severity of MDD and suicidality. qEEG was well tolerated by youth with MDD and, if further research continues to elucidate the role connectivity plays in MDD diagnosis and treatment, qEEG may be a feasible, low cost biomarker to supplement the current assessment tools available. Limitations Our study has several limitations. First, our sample was relatively small for a machine learning approach. However, this was an exploratory (or hypothesis-generating) analysis that used a rigorous cross validation framework to minimize the potential for overfitting. Second, while cross validation was implemented, validation was internal (within the same dataset). A more robust validation scheme would include external validation on an independent dataset; however, none are available at this time. Third, although qEEG has many strengths, localization and making conclusions regarding areas of the brain from EEG locations has many limitations. EEG is measuring scalp activity and only inferences can be made about deeper brain structures. Future studies with both EEG and functional imaging may help to clarify the locations of regions of interest found on EEG. Both samples were limited by the very small sample size and continue to suggest dysregulation within the DMN in adolescent depression. Additionally, they suggest locations for future study. Our prior work was the first to demonstrate differences in EEG coherence in adolescent depression. Previous work in adults has demonstrated higher connectivity in regions associated with the DMN(Ho et al., 2024 ) while functional imaging studies have demonstrated dysregulated connectivity in the DMN in adolescent depression (FIND REF). It remains unclear if the location or directional pattern of aberrant connectivity (increased vs. decreased) are meaningful regarding the pathophysiology of adolescent depression. Conclusion These findings show promise for the longitudinal investigations within this comorbidity-free, treatment naïve sample that could show how these findings endure or progress over treatment and symptom severity. Future research will evaluate whether baseline measures of qEEG can predict treatment response and will compare treatment responsive MDD youth at 16 weeks to treatment non-responders and HC. References Andersen SL, Teicher MH (2008) Stress, sensitive periods and maturational events in adolescent depression. Trends Neurosci 31(4):183–191. 10.1016/j.tins.2008.01.004 Bitsko RH, Claussen AH, Lichstein J, Black LI, Jones SE, Danielson ML, Hoenig JM, Jack D, Brody SP, Gyawali DJ, Maenner S, Warner MJ, Holland M, Perou KM, Crosby R, Blumberg AE, Avenevoli SJ, Kaminski S, J. W., Ghandour RM (2022) Mental Health Surveillance Among Children - United States, 2013–2019. MMWR Suppl 71(2):1–42. 10.15585/mmwr.su7102a1 Bittar TP, Labonté B (2021) Functional Contribution of the Medial Prefrontal Circuitry in Major Depressive Disorder and Stress-Induced Depressive-Like Behaviors. Front Behav Neurosci 15:699592. 10.3389/fnbeh.2021.699592 Breiman L (2001) Random forests. Mach Learn 45:5–32 Breukelaar IA, Antees C, Grieve SM, Foster SL, Gomes L, Williams LM, Korgaonkar MS (2017) Cognitive control network anatomy correlates with neurocognitive behavior: A longitudinal study. Hum Brain Mapp 38(2):631–643. 10.1002/hbm.23401 Busner J, Steven PD, Targum D, MD (2007) The Clinical Global Impressions Scale: Applying a Research Tool in Clinical Practice. Psychiatry (Edgmont) 4(7):28–37 Centers for Disease Control and Prevention, f. NC (2021) H. S. National Vital Statistics System, Mortality 2018–2021 on CDC WONDER Online Database, released in 2021 Chan AS, Leung WW (2006) Differentiating Autistic Children With Quantitative Encephalography: A 3-Month Longitudinal Study. J Child Neurol 21(5):392–399 Coben R, Clarke AR, Hudspeth W, Barry RJ (2008) EEG power and coherence in autistic spectrum disorder. Clin Neurophysiol 119(5):1002–1009. 10.1016/j.clinph.2008.01.013 Das A, de Los Angeles C, Menon V (2022) Electrophysiological foundations of the human default-mode network revealed by intracranial-EEG recordings during resting-state and cognition. NeuroImage 250:118927. 10.1016/j.neuroimage.2022.118927 di Michele F, Prichep L, John ER, Chabot RJ (2005) The neurophysiology of attention-deficit/hyperactivity disorder. Int J Psychophysiol 58(1):81–93. 10.1016/j.ijpsycho.2005.03.011 Dwyer JB, Stringaris A, Brent DA, Bloch MH (2020) Annual Research Review: Defining and treating pediatric treatment-resistant depression. J Child Psychol Psychiatry 61(3):312–332. 10.1111/jcpp.13202 Fitzgerald PB, Laird AR, Maller J, Daskalakis ZJ (2008) A meta-analytic study of changes in brain activation in depression. Hum Brain Mapp 29(6):683–695. 10.1002/hbm.20426 Fonseca LC, Tedrus GM, Moraes C, Vicente Machado A, Almeida MP, Oliveira DO (2008) Epileptiform abnormalities and quantitative EEG in children with attention-deficit/hyperactivity disorder. Arq Neuropsiquiatr 66(3a):462–467 Fuhrmann D, Knoll LJ, Blakemore SJ (2015) Adolescence as a Sensitive Period of Brain Development. Trends Cogn Sci 19(10):558–566. 10.1016/j.tics.2015.07.008 Gogtay N, Giedd JN, Lusk L, Hayashi KM, Greenstein D, Vaituzis AC, Nugent TF, Herman DH, Clasen LS, Toga AW, Rapoport JL, Thompson PM (2004) Dynamic mapping of human cortical development during childhood through early adulthood. Proc Natl Acad Sci USA 101(21):8174–8179. 10.1073/pnas.0402680101 Griffin A (2017) Adolescent Neurological Development and Implications for Health and Well-Being. Healthc (Basel) 5(4). 10.3390/healthcare5040062 Grinsztajn L, Oyallon E, Varoquaux G (2022) Why do tree-based models still outperform deep learning on typical tabular data? Adv Neural Inf Process Syst 35:507–520 Guan M, Liu X, Guo L, Zhang R, Tan Q, Wang H (2021) Improved Pre-attentive Processing With Occipital rTMS Treatment in Major Depressive Disorder Patients Revealed by MMN. Front Hum Neurosci 15:648816. 10.3389/fnhum.2021.648816 Hack LM, Tozzi L, Zenteno S, Olmsted AM, Hilton R, Jubeir J, Korgaonkar MS, Schatzberg AF, Yesavage JA, O'Hara R, Williams LM (2023) A Cognitive Biotype of Depression and Symptoms, Behavior Measures, Neural Circuits, and Differential Treatment Outcomes: A Prespecified Secondary Analysis of a Randomized Clinical Trial. JAMA Netw Open 6(6):e2318411. 10.1001/jamanetworkopen.2023.18411 Halfin A (2007) Depression: the benefits of early and appropriate treatment. Am J Manag Care 13(4 Suppl):S92–S97 Harder A, Nguyen TD, Pasman JA, Mosing MA, Hägg S, Lu Y (2022) Genetics of age-at-onset in major depression. Transl Psychiatry 12(1):124. 10.1038/s41398-022-01888-z Herzog DP, Wagner S, Engelmann J, Treccani G, Dreimüller N, Müller MB, Tadic A, Murck H, Lieb K (2021) Early onset of depression and treatment outcome in patients with major depressive disorder. J Psychiatr Res 139:150–158. 10.1016/j.jpsychires.2021.05.048 Ho SI, Lin IM, Hsieh JC, Yen CF (2024) EEG coherences of the default mode network among patients comorbid with major depressive disorder and anxiety symptoms. J Affect Disord 361:728–738. 10.1016/j.jad.2024.06.041 Ho TC, Connolly CG, Henje Blom E, LeWinn KZ, Strigo IA, Paulus MP, Frank G, Max JE, Wu J, Chan M, Tapert SF, Simmons AN, Yang TT (2015) Emotion-Dependent Functional Connectivity of the Default Mode Network in Adolescent Depression. Biol Psychiatry 78(9):635–646. 10.1016/j.biopsych.2014.09.002 Ismail FY, Fatemi A, Johnston MV (2017) Cerebral plasticity: Windows of opportunity in the developing brain. Eur J Paediatr Neurol 21(1):23–48. 10.1016/j.ejpn.2016.07.007 Jiao K, Xu H, Teng C, Song X, Xiao C, Fox PT, Zhang N, Wang C, Zhong Y (2020) Connectivity patterns of cognitive control network in first episode medication-naive depression and remitted depression. Behav Brain Res 379:112381. 10.1016/j.bbr.2019.112381 Johnson D, Dupuis G, Piche J, Clayborne Z, Colman I (2018) Adult mental health outcomes of adolescent depression: A systematic review. Depress Anxiety 35(8):700–716. 10.1002/da.22777 Kaiser RH, Andrews-Hanna JR, Wager TD, Pizzagalli DA (2015) Large-Scale Network Dysfunction in Major Depressive Disorder: A Meta-analysis of Resting-State Functional Connectivity. JAMA Psychiatry 72(6):603–611. 10.1001/jamapsychiatry.2015.0071 Kim JW, Lee YS, Han DH, Min KJ, Lee J, Lee K (2015) Diagnostic utility of quantitative EEG in un-medicated schizophrenia. Neurosci Lett 589:126–131. 10.1016/j.neulet.2014.12.064 Kuhn M, Silge J (2022) Tidy modeling with R: A framework for modeling in the tidyverse. O'Reilly Media, Inc. Kupfer DJ, Frank E, Phillips ML (2012) Major depressive disorder: new clinical, neurobiological, and treatment perspectives. Lancet 379(9820):1045–1055. 10.1016/s0140-6736(11)60602-8 Lee JS, Kang W, Kang Y, Kim A, Han KM, Tae WS, Ham BJ (2021) Alterations in the Occipital Cortex of Drug-Naïve Adults With Major Depressive Disorder: A Surface-Based Analysis of Surface Area and Cortical Thickness. Psychiatry Investig 18(10):1025–1033. 10.30773/pi.2021.0099 Liu Y, Pu C, Xia S, Deng D, Wang X, Li M (2022) Machine learning approaches for diagnosing depression using EEG: A review. Transl Neurosci 13(1):224–235. 10.1515/tnsci-2022-0234 Lu W, Keyes KM (2023) Major depression with co-occurring suicidal thoughts, plans, and attempts: An increasing mental health crisis in US adolescents, 2011–2020. Psychiatry Res 327:115352. 10.1016/j.psychres.2023.115352 Marsh R, Gerber AJ, Peterson BS (2008) Neuroimaging studies of normal brain development and their relevance for understanding childhood neuropsychiatric disorders. J Am Acad Child Adolesc Psychiatry 47(11):1233–1251. 10.1097/CHI.0b013e318185e703 Mayes TL, Bernstein IH, Haley CL, Kennard BD, Emslie GJ (2010) Psychometric Properties of the Children's Depression Rating Scale-Revised in Adolescents. J Child Adolesc Psychopharmacol 20(6):513–516. 10.1089/cap.2010.0063 McGorry P, Keshavan M, Goldstone S, Amminger P, Allott K, Berk M, Lavoie S, Pantelis C, Yung A, Wood S, Hickie I (2014) Biomarkers and clinical staging in psychiatry. World Psychiatry 13(3):211–223. 10.1002/wps.20144 10.1002/wps.20144 McGorry PD, Mei C (2018) Early intervention in youth mental health: progress and future directions. Evid Based Ment Health 21(4):182–184. 10.1136/ebmental-2018-300060 McVoy M, Aebi ME, Loparo K, Lytle S, Morris A, Woods N, Deyling E, Tatsuoka C, Kaffashi F, Lhatoo S, Sajatovic M (2019a) Resting-State Quantitative Electroencephalography Demonstrates Differential Connectivity in Adolescents with Major Depressive Disorder. J Child Adolesc Psychopharmacol 29(5):370–377. 10.1089/cap.2018.0166 McVoy M, Lytle S, Fulchiero E, Aebi ME, Adeleye O, Sajatovic M (2019b) A systematic review of quantitative EEG as a possible biomarker in child psychiatric disorders. Psychiatry Res. 10.1016/j.psychres.2019.07.004 Mendelson T, Tandon SD (2016) Prevention of Depression in Childhood and Adolescence. Child Adolesc Psychiatr Clin N Am 25(2):201–218. 10.1016/j.chc.2015.11.005 Moeini M, Khaleghi A, Mohammadi MR (2015) Characteristics of Alpha Band Frequency in Adolescents with Bipolar II Disorder: A Resting-State QEEG Study. Iran J Psychiatry 10(1):8–12 Mumtaz W, Ali SSA, Yasin MAM, Malik AS (2018) A machine learning framework involving EEG-based functional connectivity to diagnose major depressive disorder (MDD). Med Biol Eng Comput 56(2):233–246. 10.1007/s11517-017-1685-z Posner K, Brown GK, Stanley B, Brent DA, Yershova KV, Oquendo MA, Currier GW, Melvin GA, Greenhill L, Shen S, Mann JJ (2011) The Columbia-Suicide Severity Rating Scale: initial validity and internal consistency findings from three multisite studies with adolescents and adults. Am J Psychiatry 168(12):1266–1277. 10.1176/appi.ajp.2011.10111704 Poznanski EO, Mokros HB (1996) Children's Depression Rating Scale, Revised (CDRS-R) Manual. Western Psychological Services, Los Angeles, CA Rice F, Eyre O, Riglin L, Potter R (2017) Adolescent depression and the treatment gap. Lancet Psychiatry 4(2):86–87. 10.1016/s2215-0366(17)30004-4 Riddle MA, Ginsburg GS, Walkup J, Labelarte MJ (2002) The Pediatric Anxiety Rating Scale (PARS): Development and psychometric properties. J Am Acad Child Adolesc Psychiatry 41(9):1061–1069 Snaith P (1993) Anhedonia: a neglected symptom of psychopathology. Psychol Med 23(4):957–966 Strawn JR, Aaronson ST, Elmaadawi AZ, Schrodt GR, Holbert RC, Verdoliva S, Heart K, Demitrack MA, Croarkin PE (2020) Treatment-Resistant Depression in Adolescents: Clinical Features and Measurement of Treatment Resistance. J Child Adolesc Psychopharmacol 30(4):261–266. 10.1089/cap.2020.0008 Takahashi T, Yücel M, Lorenzetti V, Walterfang M, Kawasaki Y, Whittle S, Suzuki M, Pantelis C, Allen NB (2010) An MRI study of the superior temporal subregions in patients with current and past major depression. Prog Neuropsychopharmacol Biol Psychiatry 34(1):98–103. 10.1016/j.pnpbp.2009.10.005 Utevsky AV, Smith DV, Huettel SA (2014) Precuneus is a functional core of the default-mode network. J Neurosci 34(3):932–940. 10.1523/jneurosci.4227-13.2014 Walter HJ, Abright AR, Bukstein OG, Diamond J, Keable H, Ripperger-Suhler J, Rockhill C (2023) Clinical Practice Guideline for the Assessment and Treatment of Children and Adolescents With Major and Persistent Depressive Disorders. J Am Acad Child Adolesc Psychiatry 62(5):479–502. 10.1016/j.jaac.2022.10.001 Whitford TJ, Rennie CJ, Grieve SM, Clark CR, Gordon E, Williams LM (2007) Brain maturation in adolescence: concurrent changes in neuroanatomy and neurophysiology. Hum Brain Mapp 28(3):228–237. 10.1002/hbm.20273 Williams LM (2016) Precision psychiatry: a neural circuit taxonomy for depression and anxiety. Lancet Psychiatry 3(5):472–480. 10.1016/s2215-0366(15)00579-9 Willinger D, Häberling I, Ilioska I, Berger G, Walitza S, Brem S (2024) Weakened effective connectivity between salience network and default mode network during resting state in adolescent depression. Front Psychiatry 15:1386984. 10.3389/fpsyt.2024.1386984 Wu F, Lu Q, Kong Y, Zhang Z (2023) A Comprehensive Overview of the Role of Visual Cortex Malfunction in Depressive Disorders: Opportunities and Challenges. Neurosci Bull 39(9):1426–1438. 10.1007/s12264-023-01052-7 Yamashita A, Sakai Y, Yamada T, Yahata N, Kunimatsu A, Okada N, Itahashi T, Hashimoto R, Mizuta H, Ichikawa N, Takamura M, Okada G, Yamagata H, Harada K, Matsuo K, Tanaka SC, Kawato M, Kasai K, Kato N, Takahashi H, Okamoto Y, Yamashita O, Imamizu H (2020) Generalizable brain network markers of major depressive disorder across multiple imaging sites. PLoS Biol 18(12). 10.1371/journal.pbio.3000966 . e3000966 Yang S, Tseng KY (2022) Maturation of Corticolimbic Functional Connectivity During Sensitive Periods of Brain Development. Curr Top Behav Neurosci 53:37–53. 10.1007/7854_2021_239 Zhao P, Wang X, Wang Q, Yan R, Chattun MR, Yao Z, Lu Q (2023) Altered fractional amplitude of low-frequency fluctuations in the superior temporal gyrus: a resting-state fMRI study in anxious depression. BMC Psychiatry 23(1):847. 10.1186/s12888-023-05364-w Tables Table 1: Demographics of overall sample split by mental health condition. (Wilcoxon rank sum test for continuous and Fisher’s exact for all binary/categorical.) MDD n= 28 HC n= 27 Statistics Age Mean (SD) 15.42 (1.29) 15.29 (1.03) p= 0.67 Sex N (% Female) 24 (85.7) 19 (85.7) p= 0.21 (Fisher’s exact) Race White African American Other Asian 25 (89.3) 2 (7.14) 1 (3.57) -- 23 (82.1) 3 (11.1) 0 (0.0) 1 (3.57) χ 2 (2)= , p=0.75 SHAPS a,b total score Mean (SD), range, n 3.0 (3.4), 0-14 n= 26* 0.56 (1.0), 0-3 n=27 p=0.0004 CDRS-R c total score Mean (SD), range 63.1 (11.6), 40-84 18.5 (2.52), 17-27 p=1.3x10 -8 PARS d total score Mean (SD), range, n 15.5 (6.60), 0-28 n=26* 2.45 (4.25), 0-13 n=27 p=3.9x10 -7 Previous Med Trials** Mean (SD), range 0.73 (0.83), 0-3 -- -- Suicidal Ideation** N (%) 21 (75%) -- -- Suicidal Ideation ** Severity Mean (SD), range, n 2.43 (1.08), 1-5 n=21*** -- -- Ideation Intensity** Mean (SD), range, n 14.90 (3.57), 11-23 n=21* -- -- Previous Suicidal Attempt** N (%) 4 (14.3%) -- -- Suicidal Behavior** N (%) 1 (3.5%) -- -- a SHAPS: Snaith-Hamilton Pleasure Scale b Introduced to the study after 5 people already completed c Children’s Depression Rating Scale- Revised d Pediatric Anxiety Rating Scale *Incomplete data for 2 participants **Only reported for MDD participants *** scores only applicable for participants endorsing suicidal ideation Table 2: Model Performance Metric Estimate [SD] p-value AUC-ROC 0.65 [0.26] 0.08 AUC-PR 0.67 [0.24] 0.12 Sensitivity 0.65 [0.28] 0.01 Specificity 0.58 [0.30] 0.25 PPV 0.63 [0.24] 0.07 NPV 0.67 [0.24] 0.03 Abbreviations: AUC: Area under the curve; ROC: Receiver operating characteristic; PR: Precision-recall; PPV: Positive predictive value; NPV: Negative predictive value Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6614439","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":453440687,"identity":"2eca43f0-3800-4332-904c-63a09a5ae4dc","order_by":0,"name":"Molly McVoy","email":"data:image/png;base64,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","orcid":"","institution":"Department of Psychiatry, Case Western Reserve University School of Medicine, Cleveland, OH, USA","correspondingAuthor":true,"prefix":"","firstName":"Molly","middleName":"","lastName":"McVoy","suffix":""},{"id":453440688,"identity":"5073bba4-5e24-41f4-92c5-881bb4537e04","order_by":1,"name":"Serhiy Chumachenko","email":"","orcid":"","institution":"Department of Psychiatry and Behavioral Neuroscience, University of Cincinnati School of Medicine, Cincinnati, OH, USA","correspondingAuthor":false,"prefix":"","firstName":"Serhiy","middleName":"","lastName":"Chumachenko","suffix":""},{"id":453440761,"identity":"fd403d86-f71c-413c-892c-6462e19d9ce3","order_by":2,"name":"Maia Gersten","email":"","orcid":"","institution":"Department of Psychiatry, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA","correspondingAuthor":false,"prefix":"","firstName":"Maia","middleName":"","lastName":"Gersten","suffix":""},{"id":453440762,"identity":"972be6b3-d01d-43a0-92e1-8dfc905890c4","order_by":3,"name":"Benjamin Wade","email":"","orcid":"","institution":"Department of Psychiatry, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA","correspondingAuthor":false,"prefix":"","firstName":"Benjamin","middleName":"","lastName":"Wade","suffix":""},{"id":453440763,"identity":"cda9c159-ec53-4584-adb1-904fd0eb641f","order_by":4,"name":"Oscar Corcelles","email":"","orcid":"","institution":"Department of Psychiatry, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA","correspondingAuthor":false,"prefix":"","firstName":"Oscar","middleName":"","lastName":"Corcelles","suffix":""},{"id":453441766,"identity":"05f98490-8309-481d-bcef-ebe7e4d00e57","order_by":5,"name":"Joy Yala","email":"","orcid":"","institution":"Department of Psychiatry, Case Western Reserve University School of Medicine, Cleveland, OH, USA","correspondingAuthor":false,"prefix":"","firstName":"Joy","middleName":"","lastName":"Yala","suffix":""},{"id":453441767,"identity":"f8681822-00fe-4424-bd21-1775a8353c13","order_by":6,"name":"Mikaila Gray","email":"","orcid":"","institution":"Department of Psychiatry, Case Western Reserve University School of Medicine, Cleveland, OH, USA","correspondingAuthor":false,"prefix":"","firstName":"Mikaila","middleName":"","lastName":"Gray","suffix":""},{"id":453441834,"identity":"bfbc90c4-68af-4bc6-892b-882e7c9d9255","order_by":7,"name":"Alla Morris","email":"","orcid":"","institution":"Neurological Institute, University Hospitals Cleveland Medical Center, Cleveland, OH, USA","correspondingAuthor":false,"prefix":"","firstName":"Alla","middleName":"","lastName":"Morris","suffix":""},{"id":453443562,"identity":"d3225b9b-f04c-4674-87d1-eda1712b36ae","order_by":8,"name":"Asif Jamil","email":"","orcid":"","institution":"Department of Psychiatry, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA","correspondingAuthor":false,"prefix":"","firstName":"Asif","middleName":"","lastName":"Jamil","suffix":""},{"id":453443635,"identity":"92347d0c-d79e-44c9-b744-79a22d322904","order_by":9,"name":"Paolo Cassano","email":"","orcid":"","institution":"Division of Neuropsychiatry, Center for Anxiety and Traumatic Stress Disorders, Harvard Medical School, Boston, MA, USA","correspondingAuthor":false,"prefix":"","firstName":"Paolo","middleName":"","lastName":"Cassano","suffix":""},{"id":453443636,"identity":"412d3e41-62f6-4d71-967d-eaece4fc813e","order_by":10,"name":"Farhad Kaffashi","email":"","orcid":"","institution":"ISSACS: Institute for Smart, Secure and Connected Systems, Case Western Reserve University, Cleveland, OH, USA","correspondingAuthor":false,"prefix":"","firstName":"Farhad","middleName":"","lastName":"Kaffashi","suffix":""},{"id":453444092,"identity":"32d140d4-57e4-4822-bb83-e6c39f5357db","order_by":11,"name":"Kenneth Loparo","email":"","orcid":"","institution":"ISSACS: Institute for Smart, Secure and Connected Systems, Case Western Reserve University, Cleveland, OH, USA","correspondingAuthor":false,"prefix":"","firstName":"Kenneth","middleName":"","lastName":"Loparo","suffix":""},{"id":453444110,"identity":"4253296d-14eb-4503-8c69-077fb823c5c2","order_by":12,"name":"Farren Briggs","email":"","orcid":"","institution":"Division of Epidemiology, Department of Public Health Sciences, University of Miami Miller School of Medicine, Miami, FL, USA","correspondingAuthor":false,"prefix":"","firstName":"Farren","middleName":"","lastName":"Briggs","suffix":""},{"id":453444111,"identity":"173bf9f8-7e64-406b-9903-372caaf6f47c","order_by":13,"name":"Martha Sajatovic","email":"","orcid":"","institution":"Department of Psychiatry, Case Western Reserve University School of Medicine, Cleveland, OH, USA","correspondingAuthor":false,"prefix":"","firstName":"Martha","middleName":"","lastName":"Sajatovic","suffix":""}],"badges":[],"createdAt":"2025-05-07 18:23:39","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6614439/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6614439/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":82699497,"identity":"e6dd4691-b550-4969-ba6f-0aea49caa661","added_by":"auto","created_at":"2025-05-14 09:23:15","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":191291,"visible":true,"origin":"","legend":"\u003cp\u003eEEG Electrode Placement\u003c/p\u003e","description":"","filename":"figure1eegelectrodeplacement.png","url":"https://assets-eu.researchsquare.com/files/rs-6614439/v1/6190ae95009cb94ecc0ed02c.png"},{"id":82699475,"identity":"ff2aacec-77af-4d09-a350-19632807791e","added_by":"auto","created_at":"2025-05-14 09:23:13","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":28958,"visible":true,"origin":"","legend":"\u003cp\u003eCONSORT Diagram\u003c/p\u003e","description":"","filename":"CONSORTDiagram.png","url":"https://assets-eu.researchsquare.com/files/rs-6614439/v1/63bf08c98f511b7b5263b94c.png"},{"id":82699491,"identity":"ac8e33c9-1157-4ee9-9e31-c06d144f7da0","added_by":"auto","created_at":"2025-05-14 09:23:14","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":81174,"visible":true,"origin":"","legend":"\u003cp\u003eROC Curve\u003c/p\u003e","description":"","filename":"ROCcurve.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6614439/v1/d3e0960ae9b61296a89a3b49.jpg"},{"id":82699526,"identity":"5dee0483-6515-4281-b0e1-4ebde724c128","added_by":"auto","created_at":"2025-05-14 09:23:17","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":110373,"visible":true,"origin":"","legend":"\u003cp\u003eEEG Feature Importance within Model for MDD vs. HC\u003c/p\u003e","description":"","filename":"featureimportance.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6614439/v1/d202cb98fdbdde504ca4f242.jpg"},{"id":82699494,"identity":"6995a4ab-501f-4651-b997-2d3d16d8086e","added_by":"auto","created_at":"2025-05-14 09:23:15","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":174084,"visible":true,"origin":"","legend":"\u003cp\u003eBoxplots by diagnosis\u003c/p\u003e","description":"","filename":"Boxplotsbydiagnosis3.17.25.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6614439/v1/587760df1f94da664e1b85a6.jpg"},{"id":82699801,"identity":"1812b7e6-ee71-4aba-82f7-f8f5ea71cd16","added_by":"auto","created_at":"2025-05-14 09:23:38","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1264228,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6614439/v1/78d4c318-deab-4ee6-a007-d6ff4d7029c9.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eAssessing the Predictive Utility of Quantitative Electroencephalography Coherence in Adolescent Major Depressive Disorder: A Machine Learning Approach\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003ePsychiatric disorders are one of the leading causes of morbidity and mortality in youth in the United States.(Bitsko et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Lu and Keyes, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) Major depressive disorder (MDD) is the most common mood disorder in youth, affecting up to 20% of adolescents in the United States.(Mendelson and Tandon, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) Adolescent-onset MDD is associated with more severe and persistent symptoms than when diagnosed in adulthood(Harder et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Johnson et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), higher levels of comorbidity and poorer psychosocial functioning.(Herzog et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Strawn et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) The highly malignant effects of MDD place adolescents at risk for poor school performance, substance use disorders, self-injury, and in the worst cases, suicide.(Dwyer et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Kupfer et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Lu and Keyes, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Walter et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) Suicide was the third leading cause of death in U.S. adolescents in 2021, taking the lives of 2,343 youth.(Centers for Disease Control and Prevention, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) To put this in perspective, in the last decade, more adolescents died from suicide than from cancer, cardiovascular disease, COVID-19 birth defects, lung disease, and influenza combined.(Centers for Disease Control and Prevention, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eAdolescence marks a period of re-organization and pruning of neuronal connections and is influenced by genes, environment, physical health and other unknown factors.(Gogtay et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2004\u003c/span\u003e; Griffin, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Marsh et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Whitford et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2007\u003c/span\u003e) MDD commonly arises during this time of rapid brain development, and may change the trajectory of this development.(Andersen and Teicher, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Fuhrmann et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Ismail et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Yang and Tseng, \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) While it is a unique time of risk, adolescence is also a time of unique promise and potential.(Fuhrmann et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Ismail et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) Neuroscience research increasingly supports the notion that adolescence represents a \u0026ldquo;unique window of plasticity.\u0026rdquo;(Ismail et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) Consequently, if depression can be detected accurately and early, adolescence provides an opportunity to prevent or mitigate the devastating consequences of the disease.(Halfin, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; McGorry and Mei, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Rice et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Walter et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) Although MDD is common and potentially deadly, there are currently no objective biomarkers for this disease. Therefore, early recognition and accurate diagnosis are difficult, and rely on verbal reports from youth and parents.(McGorry et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Walter et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2023\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eWe report on a project that utilized quantitative electroencephalography (qEEG) as a diagnostic biomarker of adolescent MDD. Previous work by our team demonstrated decreased resting connectivity (qEEG coherence) in 25 MDD teens compared with 14 age and gender-match healthy controls.(McVoy et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2019a\u003c/span\u003e) This decreased connectivity was most pronounced in the right prefrontal cortex, an area of the brain associated with the default mode network (DMN). The pilot study investigated a small number of adolescents, all on multiple psychiatric medications and with significant comorbidity, thus only allowing a limited analysis.(McVoy et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2019a\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eWe now present the baseline findings from a longitudinal study comparing those with MDD compared to age- and gender-matched healthy controls (HC). The study aims to assess whether qEEG and connectivity measures, specifically coherence, reveal differences between adolescents with MDD and their non-depressed, healthy peers. In addition, this study used machine learning (ML) methods to evaluate EEG markers of MDD. Previous studies have applied machine learning methods to identify diagnostic biomarkers in adults for depression using neuroimaging data, including functional magnetic resonance imaging (fMRI)(Yamashita et al., \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and EEG.(Liu et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Mumtaz et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) To date, applications of ML to predict adolescent MDD have been developed using clinical history or electronic health records data; however, none have attempted to predict adolescent MDD using biological measures like qEEG. This study therefore presents a novel approach in the field. We hypothesized that the machine learning models will replicate the findings of our earlier research, and provide further evidence of the utility of qEEG as an intermediate biomarker of adolescent MDD.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e \u003cstrong\u003eOverall study description\u003c/strong\u003e \u003cp\u003eThis study enrolled 55 adolescents aged 14\u0026ndash;17, 28 with at least moderate MDD and 27 age- and gender-matched HC youth. MDD youth were followed over 16 weeks while HC youth had a one-time assessment visit. Clinical assessments included a diagnostic evaluation and standardized measures of depression, anhedonia, and anxiety. Following the baseline assessment, MDD youth were treated with evidence-based medication (fluoxetine or escitalopram) and a manualized brief therapy. The aim of the baseline study was to evaluate resting qEEG connectivity (coherence) in MDD cases vs. HCs.\u003c/p\u003e \u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eSubjects and recruitment:\u003c/h2\u003e \u003cp\u003eSubjects were recruited at University Hospitals Cleveland Medical Center (UHCMC), a tertiary care academic medical center, from both child psychiatry offices and general pediatrics\u0026rsquo; practices. MDD adolescents aged 13\u0026ndash;18 were included if they met DSM-5 criteria for Major Depressive Disorder (MDD) of moderate severity. Diagnosis and severity were determined through clinical evaluation by a Board Eligible/Board Certified Child Psychiatrist and the Children\u0026rsquo;s Depression Rating Scale- Revised (CDRS-R)\u0026thinsp;\u0026ge;\u0026thinsp;40. Subjects were scheduled to begin clinical treatment with escitalopram or fluoxetine for MDD. We excluded subjects with comorbid diagnoses that have reportedly shown differences on qEEG(Chan and Leung, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Coben et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; di Michele et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Fonseca et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Kim et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; McVoy et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2019b\u003c/span\u003e; Moeini et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2015\u003c/span\u003e), including attention deficit hyperactivity disorder (ADHD), bipolar spectrum disorder, autism spectrum disorder, primary psychotic disorder, active substance use disorder, post-traumatic stress disorder, and neurologic conditions including epilepsy, meningitis, history of brain surgery, shunts. In addition, treatment resistant youth, defined as having failed 2 or more trials of treatment with a selective serotonin reuptake inhibitor (SSRI) were excluded. Finally, the following medications were also exclusionary for MDD youth: benzodiazepines, atypical antipsychotics, anti-epileptics, psychostimulants, and any antidepressant medications.\u003c/p\u003e \u003cp\u003e HC adolescents were age and gender matched to MDD youth, did not meet criteria for any DSM-5 psychiatric disorder (as assessed by a Board Eligible/Board Certified Child Psychiatrist), and did not have 1st degree relatives with a history of recurrent unipolar depression, mania, hypomania or psychosis or 2nd degree relatives with a history of mania, hypomania or psychosis. The same neurologic and medication exclusion criteria as applied to MDD adolescents was applied to HC adolescents. All subjects and their caregivers provided written consent. The study was approved by the local Institutional Review Board (IRB).\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eMeasures\u003c/strong\u003e \u003cp\u003eMDD and HC youth completed the follow assessment measures at screen.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eChildren\u0026rsquo;s Depression Rating Scale-Revised (CDRS-R)\u003c/span\u003e: MDD severity was assessed with the CDRS-R.(Mayes et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Poznanski and Mokros, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e1996\u003c/span\u003e) The CDRS-R was originally derived from the Hamilton Depression Rating Scale, and is used for children aged 6\u0026ndash;17.(Poznanski and Mokros, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e1996\u003c/span\u003e) It is a 17-item scale, with items ranging from 1 to 5 or 1 to 7; total possible scores range from 17 to 113. A cut off score of \u0026le;\u0026thinsp;28 equates to minimal or no symptoms of depression, whereas a score of \u0026ge;\u0026thinsp;40 indicates clinically significant depression.\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eSnaith-Hamilton Pleasure Scale (SHAPS)\u003c/span\u003e: The SHAPS(Snaith, \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e1993\u003c/span\u003e) assesses anhedonia, or the inability to experience pleasure in normally-pleasant experiences. The SHAPS has 14 self-report Likert-scale items with the option to choose Strongly Disagree, Disagree, Agree, or Strongly Agree for each item. When scoring, each \u0026ldquo;Disagree\u0026rdquo; endorsement is given 1 point, and each \u0026ldquo;Agree\u0026rdquo; endorsement is given 0 points. Total scores range from 0\u0026ndash;14, with higher scores indicating higher anhedonia.\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ePediatric Anxiety Rating Scale (PARS)\u003c/span\u003e: Anxiety was assessed with the PARS,(Riddle et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2002\u003c/span\u003e) a 50-item symptom checklist of items grouped into categories of: Social Interactions or Performance Situations (9 items), Separation (10 items), Generalized (8 items), Specific Phobia (4 items), Physical Signs and Symptoms (13 items), and Other (6 items). Symptoms are scored on 7 dimensions of severity, using a 6-point scale (0 for none, and 1\u0026ndash;5 for minimal to extreme) for each dimension, and these are added to get a total score. Total scores range from 0 to 35, with higher scores indicating higher severity of anxiety. A score of \u0026ge;\u0026thinsp;10 indicates clinically-significant anxiety, and scores\u0026thinsp;\u0026ge;\u0026thinsp;20 indicate severe anxiety.(Riddle et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2002\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eClinical Global Impression-Severity\u003c/span\u003e (CGI-S:): CGI-S(Busner, 2007) uses clinical judgement to rate disease severity on a seven-point scale, as follows: 1\u0026thinsp;=\u0026thinsp;normal, not at all ill; 2\u0026thinsp;=\u0026thinsp;borderline mentally ill; 3\u0026thinsp;=\u0026thinsp;mildly ill; 4\u0026thinsp;=\u0026thinsp;moderately ill; 5\u0026thinsp;=\u0026thinsp;markedly ill; 6\u0026thinsp;=\u0026thinsp;severely ill; 7\u0026thinsp;=\u0026thinsp;among the most extremely ill patients.\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eColumbia Suicide Severity Scale (CSSRS)\u003c/span\u003e: Suicide risk was assessed with the CSSR-S(Posner et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2011\u003c/span\u003e), a screening tool that measures suicidality across 4 constructs: severity of ideation, intensity of ideation, behavior, and lethality. The severity subscale rates suicidal ideation on a 5-point scale: 1\u0026thinsp;=\u0026thinsp;wish to be dead, 2\u0026thinsp;=\u0026thinsp;nonspecific active suicidal thoughts, 3\u0026thinsp;=\u0026thinsp;suicidal thoughts with methods, 4\u0026thinsp;=\u0026thinsp;suicidal intent, and 5\u0026thinsp;=\u0026thinsp;suicidal intent with plan. The intensity subscale rates frequency, duration, controllability, deterrents, and reason for ideation, each on a 5-point scale. The behavior scale rates actual, aborted, and interrupted attempts, as well as preparatory behavior and nonsuicidal self-injurious behavior on a nominal scale. The lethality subscale measures actual attempts on a 6-point scale. When actual lethality scores zero, potential lethality of attempts is rated on a 3-point scale.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEEG data collection\u003c/h3\u003e\n\u003cp\u003eEEG data were acquired using a Discovery 24 EEG amplifier that has been approved by the U.S. Food and Drug Administration. Resting EEGs were recorded while subjects lay quietly with their eyes closed in a sound-attenuated room. Subjects were alerted frequently to avoid drowsiness and were instructed to remain still and inhibit blinks or eye movements during each recording period. EEGs were recorded using a 32-channel enhanced version of the International 10\u0026ndash;20 System of Electrode Placement, using the montage in Fig.\u0026nbsp;1.(McVoy et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2019a\u003c/span\u003e)\u003c/p\u003e\n\u003ch3\u003eqEEG analysis:\u003c/h3\u003e\n\u003cp\u003eWe developed a semi-automatic tool to compute the power spectral density (PSD) and coherence measures for each subject. Further details of this tool and calculation of coherence are described in the pilot paper.(McVoy et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2019a\u003c/span\u003e) The default segment length was set at 10 seconds for analysis, and the tool automatically analyzed the six eye channels (LOC-LSO, LOC-LIO, LSO-LIO, ROC-RSO, ROC-RIO, RSO-RIO) to detect segments with eye movements. PSD and coherence were then calculated for the selected EEG channels and for the delta (0.5\u0026ndash;4 Hz), theta (4\u0026ndash;8 Hz), alpha (8\u0026ndash;12 Hz), and beta (12\u0026ndash;20 Hz) frequency bands. PSD and coherence were averaged across epochs for use in data analysis, as described below.\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eData Analysis:\u003c/h2\u003e \u003cp\u003eRandom forest classifier (RF)(Breiman, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2001\u003c/span\u003e) models were used to predict individual MDD status (CDRS-R\u0026thinsp;\u0026ge;\u0026thinsp;40) using baseline EEG coherence. Previous studies have demonstrated that tree-based ensemble models like the RF model perform well on high-dimensional(Breiman, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2001\u003c/span\u003e) and tabular(Grinsztajn et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) style data structures compared to deep learning and other classical machine learning methods and are able to flexibly detect key interactions between predictors. We trained and tested RF models with 1000 underlying trees using 10-repeated, 10-fold cross validation. Our primary metric to evaluate model performance was the area under the receiver operating characteristic curve (AUC-ROC). Secondary evaluation metrics included the area under the precision-recall curve (AUC-PR), sensitivity, specificity, and positive and negative predictive values. Performance metrics were averaged across all 100 iterations of cross validation. The contribution of individual predictors was assessed using permutation importance.(Breiman, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2001\u003c/span\u003e) The significance of the model was assessed using permutation testing with B\u0026thinsp;=\u0026thinsp;1000 resamples. Models were constructed in R version 4.3.0 using tidymodels.(Kuhn and Silge, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2022\u003c/span\u003e)\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eOverall sample results\u003c/h2\u003e \u003cp\u003eDemographic and clinical variables are presented in Table\u0026nbsp;1 and a consort diagram of participants is presented in Fig.\u0026nbsp;2. Mean age of youth was 15.36 (SD\u0026thinsp;=\u0026thinsp;1.16), N\u0026thinsp;=\u0026thinsp;5855, 78% (N\u0026thinsp;=\u0026thinsp;43) female. Of the MDD participants, 27 of 28 (96%) were experiencing a current major depressive episode, and 6 (21%) had a history of recurrent depressive episodes, as assessed by the MINI-KID. Of currently active diagnoses, 32% (N\u0026thinsp;=\u0026thinsp;9) had panic disorder, 39% (N\u0026thinsp;=\u0026thinsp;11) had agoraphobia, 43% (N\u0026thinsp;=\u0026thinsp;12) had social anxiety, 18% (N\u0026thinsp;=\u0026thinsp;5) had specific phobia, 25% had generalized anxiety disorder.\u003c/p\u003e \u003cp\u003eDepression severity of the MDD participants, as measured by CDRS, was moderate, with an average of 63.1 (Range 40\u0026ndash;84). Anxiety severity, as measured by PARS, was clinically significant with an average score of 15.5, with scores ranging from zero to 28. Six participants met the cut-off of \u0026ge;\u0026thinsp;20 for severe anxiety.\u003c/p\u003e \u003cp\u003eSuicidality, as measured by the MINI-KID, was present at baseline in 39% of MDD participants (N\u0026thinsp;=\u0026thinsp;11), with 18% (N\u0026thinsp;=\u0026thinsp;5) having a lifetime attempt, and 50% having no history of suicidality. Of the 11 with current suicidality, 5 were deemed low risk, 4 medium risk, and 2 high risk. See Table\u0026nbsp;1 for details of depression and suicidality measures within the MDD sample.\u003c/p\u003e \u003cp\u003eSuicidal ideation, as measured by the CSSR-S, was present in 21 (75%) participants, with average ideation severity being 2.4 (Range 1\u0026ndash;5) and average ideation intensity being 11.0 (Range 0\u0026ndash;23). 14% of MDD (N\u0026thinsp;=\u0026thinsp;4) participants had a previous suicide attempt, 1 of which had four previous attempts. 3%(N\u0026thinsp;=\u0026thinsp;1) of MDD participants were endorsing suicidal behavior at the time of baseline assessment.\u003c/p\u003e \u003cp\u003e50% (N\u0026thinsp;=\u0026thinsp;14) of participants had no previous trials of SSRI medication. The average number of previous trials among MDD participants was 0.73 (SD\u0026thinsp;=\u0026thinsp;0.83), with a maximum observed trial of 3.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eModel Results\u003c/h3\u003e\n\u003cp\u003eRandom forest models predicted depression status with a trend-level of performance (mean AUC-ROC\u0026thinsp;=\u0026thinsp;0.65 [SD\u0026thinsp;=\u0026thinsp;0.26], p\u0026thinsp;=\u0026thinsp;0.08). Sensitivity (mean\u0026thinsp;=\u0026thinsp;0.65 [SD\u0026thinsp;=\u0026thinsp;0.28], p\u0026thinsp;=\u0026thinsp;0.01) and negative predictive value (NPV; mean\u0026thinsp;=\u0026thinsp;0.67 [SD\u0026thinsp;=\u0026thinsp;0.24], p\u0026thinsp;=\u0026thinsp;0.03) were both significant while specificity (mean\u0026thinsp;=\u0026thinsp;0.58 [SD\u0026thinsp;=\u0026thinsp;0.30], p\u0026thinsp;\u0026gt;\u0026thinsp;0.1) was non-significant and positive predictive value (PPV; 0.63 [0.24], p\u0026thinsp;=\u0026thinsp;0.07) was a trend-level of significance. Figure\u0026nbsp;3 illustrates the ROC curve for the model. See Table\u0026nbsp;2 for an outline of performance metrics. In descending order, the five most important features contributing to the prediction of MDD status were T7-P7 coherence beta, P4-O2 coherence beta, Cz-Pz coherence beta, Fz-Cz coherence delta, and Fp2-F8 coherence alpha. Figure\u0026nbsp;4 illustrates feature permutation importance for the top 10 most informative predictors. In general, lower T7-P7 coherence beta, Fz-Cz coherence delta, and Fp2-F8 coherence alpha were indicative of MDD. Conversely, higher P4-O2 coherence beta and Cz-Pz coherence beta were predictive of MDD.\u003c/p\u003e \u003cp\u003eFigure 5 illustrates boxplots of the baseline coherence measures for the top 10 most important features. Post hoc tests were conducted to examine differences between MDD and HC youth within the predictive features. As the coherence data for both populations were not normally distributed, Wilcoxon rank sum tests were performed and revealed significant differences in features T7-P7 coherence beta (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), Fz-Cz coherence delta (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), and Cz-Pz coherence beta (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). No significant differences were found in P4-O2 coherence beta (p\u0026thinsp;=\u0026thinsp;0.27) or Fp2-F8 coherence alpha (p\u0026thinsp;=\u0026thinsp;0.27). Following correction for multiple comparisons with the false discovery rate method (FDR), T7-P7 coherence beta remained significant (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eWe demonstrate a preliminary replication of our previous findings that measures of qEEG coherence differ in youth with MDD compared with HC in an expanded, comorbidity-free, treatment-na\u0026iuml;ve sample of MDD youth. Our RF classifier yielded trend-level significance in predicting adolescent MDD status using pretreatment qEEG measures. Our model had a significantly high sensitivity but non-significant specificity, suggesting that it performed better at identifying adolescents with MDD than HCs. The most informative qEEG measures in our model were biologically plausible channel pairs spanning regions often implicated in MDD including nodes of the default mode network (DMN) and cognitive control network (CCN)(Zhao et al., \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The DMN is primarily involved with emotional processing, self-referential thinking, rumination, as well as other aspects of cognition(Zhao et al., \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). It is typically found to be hyperactive in MDD, though certain regions of the DMN may exhibit hypoactivity in connection with nodes of other functional networks.(Kaiser et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) The CCN is responsible for executive functions, such as working memory, attention, and planning(Breukelaar et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), and is also primarily hypoactive in depression.(Hack et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Jiao et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2020\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eThe most predictive and significant channel pair was T7-P7, which corresponds to the superior temporal gyrus (STG) and inferior lateral occipital cortex (LOC) respectively. The STG is associated with social and emotional processing(Lee et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Takahashi et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), with previous research noting decreased activity in this region in response to sad stimuli in adult subjects with MDD.(Fitzgerald et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2008\u003c/span\u003e) It is also a primary component of the DMN. The LOC is a part of the visual association cortex that integrates visual information and has been associated with cognitive deficits or aberrant functioning in depressions.(Guan et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Wu et al., \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) Reduced activity between these two areas may reflect the diminished social and emotional processing observed in depression, particularly paying attention to and reacting to emotional visual stimuli like sad or threatening faces.(Fitzgerald et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2008\u003c/span\u003e) These deficits may be particularly relevant in adolescence, as social matters become increasingly salient.\u003c/p\u003e \u003cp\u003ePrior to correction, two other channel pairs were found to be significant and shared an electrode, Fz-Cz and Cz-Pz. However, despite this common location, there were differences in the direction of the coherence comparisons between MDD and HC adolescents. While Fz-Cz coherence was decreased in MDD youth, Cz-Pz coherence was increased. Cz roughly corresponds to the precentral gyrus, while Fz and Pz may reflect the medial prefrontal cortex (mPFC) and precuneus respectively. Both the precuneus and mPFC are important nodes of the DMN, with the precuneus being linked to complex cognitive functions such as autobiographical memory(Utevsky et al., \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2014\u003c/span\u003e) while the mPFC is associated with emotion regulation, social functioning, learning, stress response, and various cognitive functions.(Bittar and Labont\u0026eacute;, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) The precentral gyrus is more associated with the CCN rather than the DMN and is typically involved in higher order cognitive functions, like selective attention and working memory.(Williams, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) When considering its link to each of the aforementioned regions, it is possible that decreased coherence in Fz-Cz could be related to the cognitive deficits observed in depression, such as decreased cognitive flexibility, emotion regulation, and concentration, while increased coherence in Cz-Pz may reflect the dysfunction of higher order cognitive processes, for instance self-referential thinking that may become pathological rumination. Future research should explore whether the precuneus and mPFC are negatively associated with each other, particularly when considering the precentral gyrus as a central node or seed. While P4-O2 was identified as the second most important feature in our model, there were no significant differences in the coherence values between MDD and HC youth. This is surprising, given that P4 corresponds to the inferior parietal lobule (IPL), an important node of both the DMN and the CCN.\u003c/p\u003e \u003cp\u003eIn summary, we identified decreased qEEG coherence in patients with MDD versus HCs in several dyads (T7-P7, Fz-CC), but increased coherence in Cz-Pz. Notably, one of the dyad locations was empirically replicated between the two studies (Cz-Pz) for MDD youth, hinting at potential consistent localization within the DMN that could strengthen a hypothesis of dysregulation in behavioral inhibition pathways leading to initial development of MDD.(Ho et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Willinger et al., \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) However, the direction of these comparisons differed between the two studies. In our pilot, we found, decreased coherence in Cz-Pz, but conversely increased coherence in this sample. In addition, our current sample identified lower T7-P7 coherence in MDD youth as compared to HC, a finding not observed in our pilot study. However, as stated previously these locations are consistent with the DMN(Das et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) and continue to support the theory that aberrant functioning of the DMN is characteristic of adolescent MDD.\u003c/p\u003e \u003cp\u003eClinically, qEEG continues to demonstrate feasibility and practicality as a potential biomarker in youth MDD and this research confirms differences are seen in resting connectivity between youth with MDD compared to HC, even in a medication free sample. Further work continues to be needed before incorporating qEEG as part of routine clinical care, including investigating qEEG markers of connectivity as it relates to severity of MDD and suicidality. qEEG was well tolerated by youth with MDD and, if further research continues to elucidate the role connectivity plays in MDD diagnosis and treatment, qEEG may be a feasible, low cost biomarker to supplement the current assessment tools available.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eOur study has several limitations. First, our sample was relatively small for a machine learning approach. However, this was an exploratory (or hypothesis-generating) analysis that used a rigorous cross validation framework to minimize the potential for overfitting. Second, while cross validation was implemented, validation was internal (within the same dataset). A more robust validation scheme would include external validation on an independent dataset; however, none are available at this time. Third, although qEEG has many strengths, localization and making conclusions regarding areas of the brain from EEG locations has many limitations. EEG is measuring scalp activity and only inferences can be made about deeper brain structures. Future studies with both EEG and functional imaging may help to clarify the locations of regions of interest found on EEG.\u003c/p\u003e \u003cp\u003eBoth samples were limited by the very small sample size and continue to suggest dysregulation within the DMN in adolescent depression. Additionally, they suggest locations for future study. Our prior work was the first to demonstrate differences in EEG coherence in adolescent depression. Previous work in adults has demonstrated higher connectivity in regions associated with the DMN(Ho et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) while functional imaging studies have demonstrated dysregulated connectivity in the DMN in adolescent depression (FIND REF). It remains unclear if the location or directional pattern of aberrant connectivity (increased vs. decreased) are meaningful regarding the pathophysiology of adolescent depression.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThese findings show promise for the longitudinal investigations within this comorbidity-free, treatment na\u0026iuml;ve sample that could show how these findings endure or progress over treatment and symptom severity. Future research will evaluate whether baseline measures of qEEG can predict treatment response and will compare treatment responsive MDD youth at 16 weeks to treatment non-responders and HC.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAndersen SL, Teicher MH (2008) Stress, sensitive periods and maturational events in adolescent depression. Trends Neurosci 31(4):183\u0026ndash;191. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.tins.2008.01.004\u003c/span\u003e\u003cspan address=\"10.1016/j.tins.2008.01.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBitsko RH, Claussen AH, Lichstein J, Black LI, Jones SE, Danielson ML, Hoenig JM, Jack D, Brody SP, Gyawali DJ, Maenner S, Warner MJ, Holland M, Perou KM, Crosby R, Blumberg AE, Avenevoli SJ, Kaminski S, J. W., Ghandour RM (2022) Mental Health Surveillance Among Children - United States, 2013\u0026ndash;2019. MMWR Suppl 71(2):1\u0026ndash;42. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.15585/mmwr.su7102a1\u003c/span\u003e\u003cspan address=\"10.15585/mmwr.su7102a1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBittar TP, Labont\u0026eacute; B (2021) Functional Contribution of the Medial Prefrontal Circuitry in Major Depressive Disorder and Stress-Induced Depressive-Like Behaviors. Front Behav Neurosci 15:699592. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fnbeh.2021.699592\u003c/span\u003e\u003cspan address=\"10.3389/fnbeh.2021.699592\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBreiman L (2001) Random forests. Mach Learn 45:5\u0026ndash;32\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBreukelaar IA, Antees C, Grieve SM, Foster SL, Gomes L, Williams LM, Korgaonkar MS (2017) Cognitive control network anatomy correlates with neurocognitive behavior: A longitudinal study. Hum Brain Mapp 38(2):631\u0026ndash;643. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/hbm.23401\u003c/span\u003e\u003cspan address=\"10.1002/hbm.23401\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBusner J, Steven PD, Targum D, MD (2007) The Clinical Global Impressions Scale: Applying a Research Tool in Clinical Practice. Psychiatry (Edgmont) 4(7):28\u0026ndash;37\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCenters for Disease Control and Prevention, f. NC (2021) H. S. National Vital Statistics System, Mortality 2018\u0026ndash;2021 on CDC WONDER Online Database, released in 2021\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChan AS, Leung WW (2006) Differentiating Autistic Children With Quantitative Encephalography: A 3-Month Longitudinal Study. J Child Neurol 21(5):392\u0026ndash;399\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCoben R, Clarke AR, Hudspeth W, Barry RJ (2008) EEG power and coherence in autistic spectrum disorder. Clin Neurophysiol 119(5):1002\u0026ndash;1009. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.clinph.2008.01.013\u003c/span\u003e\u003cspan address=\"10.1016/j.clinph.2008.01.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDas A, de Los Angeles C, Menon V (2022) Electrophysiological foundations of the human default-mode network revealed by intracranial-EEG recordings during resting-state and cognition. NeuroImage 250:118927. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neuroimage.2022.118927\u003c/span\u003e\u003cspan address=\"10.1016/j.neuroimage.2022.118927\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003edi Michele F, Prichep L, John ER, Chabot RJ (2005) The neurophysiology of attention-deficit/hyperactivity disorder. Int J Psychophysiol 58(1):81\u0026ndash;93. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ijpsycho.2005.03.011\u003c/span\u003e\u003cspan address=\"10.1016/j.ijpsycho.2005.03.011\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDwyer JB, Stringaris A, Brent DA, Bloch MH (2020) Annual Research Review: Defining and treating pediatric treatment-resistant depression. J Child Psychol Psychiatry 61(3):312\u0026ndash;332. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/jcpp.13202\u003c/span\u003e\u003cspan address=\"10.1111/jcpp.13202\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFitzgerald PB, Laird AR, Maller J, Daskalakis ZJ (2008) A meta-analytic study of changes in brain activation in depression. Hum Brain Mapp 29(6):683\u0026ndash;695. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/hbm.20426\u003c/span\u003e\u003cspan address=\"10.1002/hbm.20426\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFonseca LC, Tedrus GM, Moraes C, Vicente Machado A, Almeida MP, Oliveira DO (2008) Epileptiform abnormalities and quantitative EEG in children with attention-deficit/hyperactivity disorder. Arq Neuropsiquiatr 66(3a):462\u0026ndash;467\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFuhrmann D, Knoll LJ, Blakemore SJ (2015) Adolescence as a Sensitive Period of Brain Development. Trends Cogn Sci 19(10):558\u0026ndash;566. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.tics.2015.07.008\u003c/span\u003e\u003cspan address=\"10.1016/j.tics.2015.07.008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGogtay N, Giedd JN, Lusk L, Hayashi KM, Greenstein D, Vaituzis AC, Nugent TF, Herman DH, Clasen LS, Toga AW, Rapoport JL, Thompson PM (2004) Dynamic mapping of human cortical development during childhood through early adulthood. Proc Natl Acad Sci USA 101(21):8174\u0026ndash;8179. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1073/pnas.0402680101\u003c/span\u003e\u003cspan address=\"10.1073/pnas.0402680101\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGriffin A (2017) Adolescent Neurological Development and Implications for Health and Well-Being. Healthc (Basel) 5(4). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/healthcare5040062\u003c/span\u003e\u003cspan address=\"10.3390/healthcare5040062\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrinsztajn L, Oyallon E, Varoquaux G (2022) Why do tree-based models still outperform deep learning on typical tabular data? Adv Neural Inf Process Syst 35:507\u0026ndash;520\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuan M, Liu X, Guo L, Zhang R, Tan Q, Wang H (2021) Improved Pre-attentive Processing With Occipital rTMS Treatment in Major Depressive Disorder Patients Revealed by MMN. Front Hum Neurosci 15:648816. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fnhum.2021.648816\u003c/span\u003e\u003cspan address=\"10.3389/fnhum.2021.648816\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHack LM, Tozzi L, Zenteno S, Olmsted AM, Hilton R, Jubeir J, Korgaonkar MS, Schatzberg AF, Yesavage JA, O'Hara R, Williams LM (2023) A Cognitive Biotype of Depression and Symptoms, Behavior Measures, Neural Circuits, and Differential Treatment Outcomes: A Prespecified Secondary Analysis of a Randomized Clinical Trial. JAMA Netw Open 6(6):e2318411. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamanetworkopen.2023.18411\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2023.18411\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHalfin A (2007) Depression: the benefits of early and appropriate treatment. Am J Manag Care 13(4 Suppl):S92\u0026ndash;S97\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarder A, Nguyen TD, Pasman JA, Mosing MA, H\u0026auml;gg S, Lu Y (2022) Genetics of age-at-onset in major depression. Transl Psychiatry 12(1):124. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41398-022-01888-z\u003c/span\u003e\u003cspan address=\"10.1038/s41398-022-01888-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHerzog DP, Wagner S, Engelmann J, Treccani G, Dreim\u0026uuml;ller N, M\u0026uuml;ller MB, Tadic A, Murck H, Lieb K (2021) Early onset of depression and treatment outcome in patients with major depressive disorder. J Psychiatr Res 139:150\u0026ndash;158. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jpsychires.2021.05.048\u003c/span\u003e\u003cspan address=\"10.1016/j.jpsychires.2021.05.048\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHo SI, Lin IM, Hsieh JC, Yen CF (2024) EEG coherences of the default mode network among patients comorbid with major depressive disorder and anxiety symptoms. J Affect Disord 361:728\u0026ndash;738. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jad.2024.06.041\u003c/span\u003e\u003cspan address=\"10.1016/j.jad.2024.06.041\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHo TC, Connolly CG, Henje Blom E, LeWinn KZ, Strigo IA, Paulus MP, Frank G, Max JE, Wu J, Chan M, Tapert SF, Simmons AN, Yang TT (2015) Emotion-Dependent Functional Connectivity of the Default Mode Network in Adolescent Depression. Biol Psychiatry 78(9):635\u0026ndash;646. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.biopsych.2014.09.002\u003c/span\u003e\u003cspan address=\"10.1016/j.biopsych.2014.09.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIsmail FY, Fatemi A, Johnston MV (2017) Cerebral plasticity: Windows of opportunity in the developing brain. Eur J Paediatr Neurol 21(1):23\u0026ndash;48. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ejpn.2016.07.007\u003c/span\u003e\u003cspan address=\"10.1016/j.ejpn.2016.07.007\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiao K, Xu H, Teng C, Song X, Xiao C, Fox PT, Zhang N, Wang C, Zhong Y (2020) Connectivity patterns of cognitive control network in first episode medication-naive depression and remitted depression. Behav Brain Res 379:112381. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.bbr.2019.112381\u003c/span\u003e\u003cspan address=\"10.1016/j.bbr.2019.112381\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJohnson D, Dupuis G, Piche J, Clayborne Z, Colman I (2018) Adult mental health outcomes of adolescent depression: A systematic review. Depress Anxiety 35(8):700\u0026ndash;716. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/da.22777\u003c/span\u003e\u003cspan address=\"10.1002/da.22777\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKaiser RH, Andrews-Hanna JR, Wager TD, Pizzagalli DA (2015) Large-Scale Network Dysfunction in Major Depressive Disorder: A Meta-analysis of Resting-State Functional Connectivity. JAMA Psychiatry 72(6):603\u0026ndash;611. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamapsychiatry.2015.0071\u003c/span\u003e\u003cspan address=\"10.1001/jamapsychiatry.2015.0071\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim JW, Lee YS, Han DH, Min KJ, Lee J, Lee K (2015) Diagnostic utility of quantitative EEG in un-medicated schizophrenia. Neurosci Lett 589:126\u0026ndash;131. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neulet.2014.12.064\u003c/span\u003e\u003cspan address=\"10.1016/j.neulet.2014.12.064\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKuhn M, Silge J (2022) Tidy modeling with R: A framework for modeling in the tidyverse. O'Reilly Media, Inc.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKupfer DJ, Frank E, Phillips ML (2012) Major depressive disorder: new clinical, neurobiological, and treatment perspectives. Lancet 379(9820):1045\u0026ndash;1055. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s0140-6736(11)60602-8\u003c/span\u003e\u003cspan address=\"10.1016/s0140-6736(11)60602-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee JS, Kang W, Kang Y, Kim A, Han KM, Tae WS, Ham BJ (2021) Alterations in the Occipital Cortex of Drug-Na\u0026iuml;ve Adults With Major Depressive Disorder: A Surface-Based Analysis of Surface Area and Cortical Thickness. Psychiatry Investig 18(10):1025\u0026ndash;1033. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.30773/pi.2021.0099\u003c/span\u003e\u003cspan address=\"10.30773/pi.2021.0099\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Y, Pu C, Xia S, Deng D, Wang X, Li M (2022) Machine learning approaches for diagnosing depression using EEG: A review. Transl Neurosci 13(1):224\u0026ndash;235. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1515/tnsci-2022-0234\u003c/span\u003e\u003cspan address=\"10.1515/tnsci-2022-0234\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu W, Keyes KM (2023) Major depression with co-occurring suicidal thoughts, plans, and attempts: An increasing mental health crisis in US adolescents, 2011\u0026ndash;2020. Psychiatry Res 327:115352. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.psychres.2023.115352\u003c/span\u003e\u003cspan address=\"10.1016/j.psychres.2023.115352\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarsh R, Gerber AJ, Peterson BS (2008) Neuroimaging studies of normal brain development and their relevance for understanding childhood neuropsychiatric disorders. J Am Acad Child Adolesc Psychiatry 47(11):1233\u0026ndash;1251. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/CHI.0b013e318185e703\u003c/span\u003e\u003cspan address=\"10.1097/CHI.0b013e318185e703\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMayes TL, Bernstein IH, Haley CL, Kennard BD, Emslie GJ (2010) Psychometric Properties of the Children's Depression Rating Scale-Revised in Adolescents. J Child Adolesc Psychopharmacol 20(6):513\u0026ndash;516. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1089/cap.2010.0063\u003c/span\u003e\u003cspan address=\"10.1089/cap.2010.0063\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcGorry P, Keshavan M, Goldstone S, Amminger P, Allott K, Berk M, Lavoie S, Pantelis C, Yung A, Wood S, Hickie I (2014) Biomarkers and clinical staging in psychiatry. World Psychiatry 13(3):211\u0026ndash;223. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/wps.20144 10.1002/wps.20144\u003c/span\u003e\u003cspan address=\"10.1002/wps.20144 10.1002/wps.20144\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcGorry PD, Mei C (2018) Early intervention in youth mental health: progress and future directions. Evid Based Ment Health 21(4):182\u0026ndash;184. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/ebmental-2018-300060\u003c/span\u003e\u003cspan address=\"10.1136/ebmental-2018-300060\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcVoy M, Aebi ME, Loparo K, Lytle S, Morris A, Woods N, Deyling E, Tatsuoka C, Kaffashi F, Lhatoo S, Sajatovic M (2019a) Resting-State Quantitative Electroencephalography Demonstrates Differential Connectivity in Adolescents with Major Depressive Disorder. J Child Adolesc Psychopharmacol 29(5):370\u0026ndash;377. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1089/cap.2018.0166\u003c/span\u003e\u003cspan address=\"10.1089/cap.2018.0166\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcVoy M, Lytle S, Fulchiero E, Aebi ME, Adeleye O, Sajatovic M (2019b) A systematic review of quantitative EEG as a possible biomarker in child psychiatric disorders. Psychiatry Res. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.psychres.2019.07.004\u003c/span\u003e\u003cspan address=\"10.1016/j.psychres.2019.07.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMendelson T, Tandon SD (2016) Prevention of Depression in Childhood and Adolescence. Child Adolesc Psychiatr Clin N Am 25(2):201\u0026ndash;218. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.chc.2015.11.005\u003c/span\u003e\u003cspan address=\"10.1016/j.chc.2015.11.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoeini M, Khaleghi A, Mohammadi MR (2015) Characteristics of Alpha Band Frequency in Adolescents with Bipolar II Disorder: A Resting-State QEEG Study. Iran J Psychiatry 10(1):8\u0026ndash;12\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMumtaz W, Ali SSA, Yasin MAM, Malik AS (2018) A machine learning framework involving EEG-based functional connectivity to diagnose major depressive disorder (MDD). Med Biol Eng Comput 56(2):233\u0026ndash;246. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11517-017-1685-z\u003c/span\u003e\u003cspan address=\"10.1007/s11517-017-1685-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePosner K, Brown GK, Stanley B, Brent DA, Yershova KV, Oquendo MA, Currier GW, Melvin GA, Greenhill L, Shen S, Mann JJ (2011) The Columbia-Suicide Severity Rating Scale: initial validity and internal consistency findings from three multisite studies with adolescents and adults. Am J Psychiatry 168(12):1266\u0026ndash;1277. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1176/appi.ajp.2011.10111704\u003c/span\u003e\u003cspan address=\"10.1176/appi.ajp.2011.10111704\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoznanski EO, Mokros HB (1996) Children's Depression Rating Scale, Revised (CDRS-R) Manual. Western Psychological Services, Los Angeles, CA\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRice F, Eyre O, Riglin L, Potter R (2017) Adolescent depression and the treatment gap. Lancet Psychiatry 4(2):86\u0026ndash;87. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s2215-0366(17)30004-4\u003c/span\u003e\u003cspan address=\"10.1016/s2215-0366(17)30004-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRiddle MA, Ginsburg GS, Walkup J, Labelarte MJ (2002) The Pediatric Anxiety Rating Scale (PARS): Development and psychometric properties. J Am Acad Child Adolesc Psychiatry 41(9):1061\u0026ndash;1069\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSnaith P (1993) Anhedonia: a neglected symptom of psychopathology. Psychol Med 23(4):957\u0026ndash;966\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStrawn JR, Aaronson ST, Elmaadawi AZ, Schrodt GR, Holbert RC, Verdoliva S, Heart K, Demitrack MA, Croarkin PE (2020) Treatment-Resistant Depression in Adolescents: Clinical Features and Measurement of Treatment Resistance. J Child Adolesc Psychopharmacol 30(4):261\u0026ndash;266. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1089/cap.2020.0008\u003c/span\u003e\u003cspan address=\"10.1089/cap.2020.0008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTakahashi T, Y\u0026uuml;cel M, Lorenzetti V, Walterfang M, Kawasaki Y, Whittle S, Suzuki M, Pantelis C, Allen NB (2010) An MRI study of the superior temporal subregions in patients with current and past major depression. Prog Neuropsychopharmacol Biol Psychiatry 34(1):98\u0026ndash;103. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.pnpbp.2009.10.005\u003c/span\u003e\u003cspan address=\"10.1016/j.pnpbp.2009.10.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUtevsky AV, Smith DV, Huettel SA (2014) Precuneus is a functional core of the default-mode network. J Neurosci 34(3):932\u0026ndash;940. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1523/jneurosci.4227-13.2014\u003c/span\u003e\u003cspan address=\"10.1523/jneurosci.4227-13.2014\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWalter HJ, Abright AR, Bukstein OG, Diamond J, Keable H, Ripperger-Suhler J, Rockhill C (2023) Clinical Practice Guideline for the Assessment and Treatment of Children and Adolescents With Major and Persistent Depressive Disorders. J Am Acad Child Adolesc Psychiatry 62(5):479\u0026ndash;502. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jaac.2022.10.001\u003c/span\u003e\u003cspan address=\"10.1016/j.jaac.2022.10.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWhitford TJ, Rennie CJ, Grieve SM, Clark CR, Gordon E, Williams LM (2007) Brain maturation in adolescence: concurrent changes in neuroanatomy and neurophysiology. Hum Brain Mapp 28(3):228\u0026ndash;237. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/hbm.20273\u003c/span\u003e\u003cspan address=\"10.1002/hbm.20273\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilliams LM (2016) Precision psychiatry: a neural circuit taxonomy for depression and anxiety. Lancet Psychiatry 3(5):472\u0026ndash;480. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s2215-0366(15)00579-9\u003c/span\u003e\u003cspan address=\"10.1016/s2215-0366(15)00579-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWillinger D, H\u0026auml;berling I, Ilioska I, Berger G, Walitza S, Brem S (2024) Weakened effective connectivity between salience network and default mode network during resting state in adolescent depression. Front Psychiatry 15:1386984. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fpsyt.2024.1386984\u003c/span\u003e\u003cspan address=\"10.3389/fpsyt.2024.1386984\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu F, Lu Q, Kong Y, Zhang Z (2023) A Comprehensive Overview of the Role of Visual Cortex Malfunction in Depressive Disorders: Opportunities and Challenges. Neurosci Bull 39(9):1426\u0026ndash;1438. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s12264-023-01052-7\u003c/span\u003e\u003cspan address=\"10.1007/s12264-023-01052-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYamashita A, Sakai Y, Yamada T, Yahata N, Kunimatsu A, Okada N, Itahashi T, Hashimoto R, Mizuta H, Ichikawa N, Takamura M, Okada G, Yamagata H, Harada K, Matsuo K, Tanaka SC, Kawato M, Kasai K, Kato N, Takahashi H, Okamoto Y, Yamashita O, Imamizu H (2020) Generalizable brain network markers of major depressive disorder across multiple imaging sites. PLoS Biol 18(12). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pbio.3000966\u003c/span\u003e\u003cspan address=\"10.1371/journal.pbio.3000966\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. e3000966\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang S, Tseng KY (2022) Maturation of Corticolimbic Functional Connectivity During Sensitive Periods of Brain Development. Curr Top Behav Neurosci 53:37\u0026ndash;53. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/7854_2021_239\u003c/span\u003e\u003cspan address=\"10.1007/7854_2021_239\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao P, Wang X, Wang Q, Yan R, Chattun MR, Yao Z, Lu Q (2023) Altered fractional amplitude of low-frequency fluctuations in the superior temporal gyrus: a resting-state fMRI study in anxious depression. BMC Psychiatry 23(1):847. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12888-023-05364-w\u003c/span\u003e\u003cspan address=\"10.1186/s12888-023-05364-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1:\u003c/strong\u003e Demographics of overall sample split by mental health condition. (Wilcoxon rank sum test for continuous and Fisher’s exact for all binary/categorical.)\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"619\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eMDD\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003en= 28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eHC\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003en= 27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eStatistics\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAge Mean (SD)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e15.42 (1.29)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e15.29 (1.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003ep= 0.67\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSex N (% Female)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e24 (85.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e19 (85.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003ep= 0.21 (Fisher’s exact)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRace\u003c/p\u003e\n \u003cp\u003eWhite\u003c/p\u003e\n \u003cp\u003eAfrican American\u003c/p\u003e\n \u003cp\u003eOther\u003c/p\u003e\n \u003cp\u003eAsian\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e25 (89.3)\u003c/p\u003e\n \u003cp\u003e2 (7.14)\u003c/p\u003e\n \u003cp\u003e1 (3.57)\u003c/p\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e23 (82.1)\u003c/p\u003e\n \u003cp\u003e3 (11.1)\u003c/p\u003e\n \u003cp\u003e0 (0.0)\u003c/p\u003e\n \u003cp\u003e1 (3.57)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eχ\u003csup\u003e2\u003c/sup\u003e(2)= , p=0.75\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSHAPS\u003csup\u003ea,b\u003c/sup\u003e total score Mean (SD), range, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3.0 (3.4), 0-14\u003cbr\u003e\u0026nbsp;n= 26*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.56 (1.0), 0-3\u003cbr\u003e\u0026nbsp;n=27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003ep=0.0004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCDRS-R\u003csup\u003ec\u003c/sup\u003e total score Mean (SD), range\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e63.1 (11.6), 40-84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e18.5 (2.52), 17-27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003ep=1.3x10\u003csup\u003e-8\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePARS\u003csup\u003ed\u003c/sup\u003e total score Mean (SD), range, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e15.5 (6.60), 0-28\u003cbr\u003e\u0026nbsp;n=26*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.45 (4.25), 0-13\u003cbr\u003e\u0026nbsp;n=27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003ep=3.9x10\u003csup\u003e-7\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePrevious Med Trials** Mean (SD), range\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.73 (0.83), 0-3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e--\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSuicidal Ideation** N (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e21 (75%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSuicidal Ideation ** Severity Mean (SD), range, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.43 (1.08), 1-5\u003cbr\u003e\u0026nbsp;n=21***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIdeation Intensity** Mean (SD), range, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e14.90 (3.57), 11-23\u003cbr\u003e\u0026nbsp;n=21*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePrevious Suicidal Attempt** N (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4 (14.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSuicidal Behavior** N (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1 (3.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003csup\u003ea\u0026nbsp;\u003c/sup\u003eSHAPS: Snaith-Hamilton Pleasure Scale\u003c/p\u003e\n\u003cp\u003e\u003csup\u003eb\u0026nbsp;\u003c/sup\u003eIntroduced to the study after 5 people already completed\u003c/p\u003e\n\u003cp\u003e\u003csup\u003ec\u0026nbsp;\u003c/sup\u003eChildren’s Depression Rating Scale- Revised\u003c/p\u003e\n\u003cp\u003e\u003csup\u003ed\u0026nbsp;\u003c/sup\u003ePediatric Anxiety Rating Scale\u003c/p\u003e\n\u003cp\u003e*Incomplete data for 2 participants\u003c/p\u003e\n\u003cp\u003e**Only reported for MDD participants\u003c/p\u003e\n\u003cp\u003e*** scores only applicable for participants endorsing suicidal ideation\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2:\u003c/strong\u003e Model Performance\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"318\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003eMetric\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003eEstimate [SD]\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003ep-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eAUC-ROC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.65 [0.26]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.08\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eAUC-PR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.67 [0.24]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.12\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eSensitivity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.65 [0.28]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eSpecificity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.58 [0.30]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.25\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003ePPV\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.63 [0.24]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.07\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eNPV\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.67 [0.24]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"bottom\"\u003e\n \u003cp\u003eAbbreviations:\u0026nbsp;\u003c/p\u003e\n \u003cp\u003eAUC: Area under the curve; ROC: Receiver operating characteristic; PR: Precision-recall; PPV: Positive predictive value; NPV: Negative predictive value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"}],"fulltextSource":"","fullText":"","funders":[{"identity":"32170f71-c90c-43c9-a3a1-2a27dbcf624b","identifier":"10.13039/100006108","name":"National Center for Advancing Translational Sciences","awardNumber":"UM1TR004528","order_by":0}],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Case Western Reserve University","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"major depressive disorder, adolescent, qEEG, quantitative electroencephologram, biomarker, connectivity","lastPublishedDoi":"10.21203/rs.3.rs-6614439/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6614439/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003eImproving early recognition and accurate diagnosis of Major Depressive Disorder (MDD) in childhood is a pressing concern. Quantitative electroencephalogram (qEEG) may be an effective, non-invasive diagnostic biomarker for MDD. Prior work by our team demonstrated decreased resting connectivity, as measured by qEEG coherence, in a heterogenous group of adolescents with MDD compared to age and gender matched health controls. This study explored qEEG coherence as a predictor of MDD diagnosis in a prospective, longitudinal sample of medication-free, adolescents with MDD vs healthy controls (HCs).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eTwenty-eight adolescents with MDD (Children’s Depression Rating Scale score ≥ 40), and twenty-seven age and gender matched HCs, (age 14-17, 78% female) received a baseline resting 32-channel EEG. Brain-wide coherence between channel pairs was calculated for the frequency bands (alpha, beta, theta, and delta) and compared between MDD youth and HC. Random forest classifiers were used to predict individual MDD status using baseline qEEG coherence. Models were trained and tested using 10-repeated, 10-fold cross validation and performance was evaluated with the area under the receiver operating characteristic curve (AUC-ROC). The contribution of individual predictors was assessed using permutation importance. Model significance was assessed using permutation testing (B=1000 resamples).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003eRandom forest models predicted depression status with a trend-level of significance (mean AUC-ROC=0.65, p=0.08). Among the most predictive channel pairs, adolescent MDD was characterized by lower coherence in T7-P7 (p\u0026lt;0.05), Fz-Cz, and Fp2-F8 as well as higher coherence in P4-O2 and Cz-Pz.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions: \u003c/strong\u003eThis study provides preliminary evidence that multivariate patterns of qEEG may inform diagnosis of adolescent MDD. Specific aberrant patterns of coherence within the default mode network and cognitive control network were characteristic of adolescent MDD. Ongoing work will seek to replicate these findings in a larger cohort.\u003c/p\u003e","manuscriptTitle":"Assessing the Predictive Utility of Quantitative Electroencephalography Coherence in Adolescent Major Depressive Disorder: A Machine Learning Approach","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-14 09:23:07","doi":"10.21203/rs.3.rs-6614439/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0062d33c-db2d-46b3-ad1b-3ffb9eca719e","owner":[],"postedDate":"May 14th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":48222627,"name":"Psychiatry"}],"tags":[],"updatedAt":"2025-05-14T09:23:07+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-14 09:23:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6614439","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6614439","identity":"rs-6614439","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}