T4 ian goodyer_escap_lecture depression
1 like954 views
Common mental illnesses often emerge between 10 and 30 years of age. Endophenotypes are likely formed by the first two decades of life, while activation processes may occur proximal to illness emergence. Depressive symptoms vary in severity and presentation depending on age and sample characteristics. Biomarkers like cognitive tests and gene-environment interactions can help identify individuals at risk of developing depression.
T4 ian goodyer_escap_lecture depression
- 1. The Emergence Of Clinical Depressions In The Human Life Cycle Professor Ian M Goodyer Department of Psychiatry University of Cambridge
- 2. • The Scope and Characteristics of the Problem. • Depression Severity and Psychotic Experiences. • Mathematical approaches to phenotypes. • Discovering Biomarkers. • The Maturing brain. This Lecture
- 3. Depressed mood Irritability/anger Pervasive anhedonia Weight/appetite disturbance Sleep disturbance Psychomotor disturbance Fatigue, lack of energy, tiredness Negative Self-perceptions Executive Cognitive disturbance Suicide Descriptive Psychopathology Unipolar Major Depression = 1 = 1 mood + 4 (or more) others A Symptoms (1 only) B Symptoms (>=4)+ High Reliability but Low Validity
- 4. The Emergence Of The Depressions: The Correlates of Age
- 5. Prevalence Of Youth Diagnostic Depressions In The First 2 Decades Of Life • <1% in Pre-‐pubertal Children: B=G. • 3%-‐6% in post-‐pubertal adolescents: G >B 2:1. • ~70% -‐> premorbid non specific difficulMes. • Higher symptoms -‐> more severity -‐> lower T response. • Clinical typology is top down and heterogeneous. • High reliability but low validity.
- 6. Common mental illnesses are emergent between 10 and 30 years. Endophenotypes likely to be formed by the first two decades of life. In contrast activation processes may occur proximal to illness emergence Developmental Epidemiology of disease
- 7. The Emergence Of The Depressions: Illness Severity
- 8. Depressed mood Irritability/anger Pervasive anhedonia Weight/appetite disturbance Sleep disturbance Psychomotor disturbance Fatigue, lack of energy, tiredness Negative Self-perceptions Executive Cognitive disturbance Suicide Descriptive Psychopathology Unipolar Major Depression = 1 = 1 mood + 4 (or more) others A Symptoms (1 only) B Symptoms (>=4)+ How many possible permuta2ons are there ? >1,000 How many occur -‐ not known High Reliability BUT low validity
- 9. Latent Trait for depression Low High Item Location One Item at medium strength on the trait Item Discrimination between persons in different regions on the latent continuum p (guess) Item Response Theory: A mathematical approach that accounts for location, discrimination and chance IndividualVariation :
- 10. -3 0 +3 Latent trait for depression Common variance between items reveals the latent trait Unique variance of the item reveals its singular importance Tears hopelesness insomnia retardation anhedonia Latent Trait for Depression: Construct Validity 1 variaMon 0 Unique variance for each item Behaviour in the adolescent populaMon Common Uncommon
- 11. • The y-‐axis is the probability that the SMFQ symptom is endorsed. • All items funcMon at more or less the same level on the latent trait. • All items are located towards the more severe end—to the right of the figures. • The probability of endorsing any item is very low. Sharp et al (2006) J Abnorm Child Psychol ;34(3):379-‐91. -‐3 0 +3 -‐3 0 + 3 -‐3 0 +3 -‐3 0 +3 Latent trait for depression 1 V a r i a t i o n 0 IRT For 657 Children 7-11 years for self reported depressive symptoms across the latent trait
- 12. IRT Model gives 2 dimensions for depressive symptoms Latent trait for depression and 2nd for maturation. Atypical depressive items load on the 2nd only IRT For 2,307 teens aged 11-18 years for depressive symptoms across the latent traits Cole D et al (2012 ) J Abnorm Psychol. 121(4):838-51
- 13. Summary • The magnitude of individual item response for contributing to a clinical state vary with age. • The importance of items varies with sample type. • Metabolic effects during adolescence account for weight gain and appetite increases. • Clinical diagnostic markers for primary care and hospital practice are likely to be different.
- 14. Revealing Structure of Clinical Phenotypes
- 15. Using 33 item MFQ and 28 item RMAS D: Depression. A: anxiety. W:Worrying. S: Somatic symptoms. G :General distress factor. Sp1-3: Specific factors. IRT : Hierarchical Bi-Factor Modeling 1159 respondents aged 14 yrs. Sex effects tested (ns)= set to zero. 8% Any Dep; 6% Any Anx by 14 yrs. Incl. correlated errors >0.6 considerably improved the fit. NS effects of instrument/method Brodbeck et al (2011) BMC Psychiatry 11:191
- 16. Brodbeck et al (2014) J Affect Disord.52-154:299-305 Bifactor Model and DiagnosticTypologies Distress Worry Hopeless SomaMc DSM Diagnoses Currently & 3 years later
- 17. Depressions and Psychotic Experience
- 18. Psychotic Experiences (PE) and Depression • PE are common in the general population (3%=5%). • PE and MDD are co-occurring. • Share the same risk factors. • Conceptual, clinical and causal links exist. • No clear cut validity for distinction in clinical typology. • Excluded from diagnostic criteria for MDD.
- 19. PsychoMc Experiences (PE), Depressive And Anxiety (D&A) Symptoms Stochl , J., et al (2015) Psychological Medicine, 45 :07, 1483-‐1493
- 20. Location Of Psychotic Experiences Relative To Depressive And Anxiety Symptoms Stochl , J., et al (2015) Psychological Medicine, 45 :07,1483-‐1493
- 21. Summary Bifactor models reveal a common general latent trait that links behaviourally different items. This is likely to account for covariance at the factor level; comorbidity at the clinical level. Need a much greater scientific understanding of the behavioural repertoire in the ‘natural world’ First step in creating new valid clinical typologies.
- 22. Discovering Biomarkers in the Adolescent Population: “a biologic feature that can be used to measure the presence or progress of disease or the effects of treatment.”
- 23. Gene-‐Environment PopulaMon Markers For The Presence of Depressions
- 24. MulMlevel Gene–environments And Symptoms A Longitudinal PerspecMve 5HTTLPR ll Child maltreatment Yes Symptoms CogniMon Child maltreatment No Symptoms CogniMon ls Child Maltreatment Yes Symptoms CogniMon Child maltreatment No Symptoms CogniMon ss Child Maltreatment No Symptoms CogniMon Chid Maltreatment Yes Symptoms CogniMon MulMple SEM tests the effects simultaneously via GLM regression whilst controlling the covariance for each equaMon.
- 25. Moderation-mediation of Cognition and Symptoms in 277 16-17 yr olds Probabilistic Reversal More errors (p=0.004) & switching in SS (CA +ve) (p=0.02). Affective G-NG SS/CA+ve ->more commission errors on the neutral task(p=0.01) AGN Neutral commission errors (OR=1.82, p=0.006) Anxiety Depressive Disorders SS(CA+ve) ->higher mean depression and anxiety scores at 14 yrs. Longitudinal prediction from cognition to symptoms at 18-19 years A B C D Owens et al (2012) PLoS One. 7(11):e48482 Depression Anxiety
- 26. Psychoendocrine PopulaMon Markers Depressive CogniMons and Clinical Disorders
- 27. Psychoendocrine Subgroups In 2 Distinct Adolescent Community Populations in Cambridge Depression symptoms and morning cor2sol level from 3 assessments over 2me Low D scores Low Cor2sol High D scores Low Cor2sol Low D Scores High Cor2sol High D Scores High Cor2sol Define subgroups using latent class analysis Classes created from a discovery sample n=666 with both measures at0,8 and 12 months Replicated in a 2nd sample , n=1198 with D measures at 0, 18,36 months but Cort at only
- 28. Clinical Depression Cases By Class 0 20 40 60 80 100 120 Class 1 Class 2 Class 3 Class 4 Not MD (n=1,524) MD (n= 207) • Each Sta2s2cally derived sub type or class has a % of depressed cases by 17 years of age. • Theore2cally these depressed cases from each sub-‐type will have different mechanisms Accoun2ng for the emergence of depressions. Owens M et al. PNAS 2014;111:3638-3643 Low D+C High C only High D only High D+C N= 539 475 421 296
- 29. The odds ratios for MD in each class by sex Owens M et al. PNAS 2014;111:3638-3643 ©2014 by National Academy of Sciences The odds ratios for MD in each class by sex. The reference group is class 1 (n = 539). Adjusted for cohort, age, and pubertal status.
- 30. LCA 4 classes and overgeneral memory (OGM) Class 4 > OGM responses than all other classes (4>1, P < 0.01),( 4>2, P < 0.001) and(4>3, P = 0.01). No sex × classes interaction (P = 0.83). N=660 Owens M et al. PNAS 2014;111:3638-3643
- 31. Summary • 5HTTLPR ‘s’ carriers + child maltreatment at risk for high anxiety and depressive symptoms . • Impaired bottom up emotion processing and/or difficulties in top down ‘learning through uncertainty’. • Dual processing cognitive deficits hypothesis. • High depression/distress traits + high morning trait cortisol defines a very high risk population sub type of adolescents. • Characterised by impairments in autobiographical memory for both sexes and in boys only for clinical depressions. • Corticoid mediated cognitive hypothesis.
- 32. Timing And The Developing Human Brain: ImplicaMons Of Timing Of Experiences And The Emergence Of Depression
- 33. Brain development proceeds in stages that vary across regions. Hippocampal volumes are 85% of adult values by adrenarche. Comparatively occurs in all mammalian species. Rates similar across species including the onset of puberty and higher-level cognition. Gray Matter Changes Cortical And Subcortical Brain Regions: 6-23 Years Based on Data from Jay Giedd: published in Andersen and Teicher 2008 TINS
- 34. • Meta-‐analysis of 23 studies: GM reducMon in bilateral rostral ACC . • ReducMon in rostral ACC the most consistent. • ReducMons in other regions within fronto-‐subcorMcal and limbic regions was less consistent. • Related posiMvely to illness duraMon. • Chronic/persistent MDD has a deleterious and perhaps focal effect on brain structure Grey Matter Reduction And Affective Disorders Meta-Analysis Findings
- 35. • X-‐SecMonal structural neuroimaging. • 109 MDD 36 healthy controls Case-‐control comparison. • F>M (3:1) ; 11-‐17 years. • GMV in ACC and across the whole-‐brain. The Depressed But Still Developing Brain
- 36. Main effect of age on GMV: controls>MDD. Age differences are dissimilar between MDD and controls. Age & Symptom Correlates of GMV in the ACC. Hagan et al 2015 , Neuroimage:Clinical
- 37. Opposite to ACC : MDD > CON. MDD only: GMV in thalamus (not ACC) 1/symptoms. Age and Depressive Symptom GMV decreases in the Thalamus Unpublished Results
- 38. • Dissimilar age-‐related and symptom-‐sensiMve parerns of GMV differences compared with controls. • The thalamus and ACC may comprise disMncMve neural markers for detecMng these effects in youth. • CriMcal to disaggregate antecedent neural vulnerabiliMes for MDD from the effects of MDD on the developing brain. Summary The Depressed And Developing Adolescent Brain
- 39. The Neural Maturation Gap: Understanding The Importance Of Brain Development Observation Early consolidation of limbic-sub-cortical reward processing networks. Later consolidation of neocortical control networks. Spike in drug use, psychotic and mood disorders in the neural maturation gap. Hypothesis Increased incidence of psychopathology in adolescence associated with different developmental rates for limbic and prefrontal systems. Proposed Mechanism Variation in rate of myelination of long distance cortico-cortical tracts predicts developmental reconfiguration of large scale brain networks. Experience dependent synaptic plasticity and pruning of inactive connections are other plausible mechanisms. IncidenceNeuraldevelopment Limbic System PFC Drug use, Psychosis, Depressions
- 40. Faculty Peter Jones Ed Bullmore Peter Fonagy John Suckling Tim Croudace Paul Fletcher Barbara Sahakian Research Staff Matt Owens Michelle St Clair Jeannette Brodbeck Cindy Hagan Kirstie Whittaker Anne Laura van Harmelen 16 Graduate Assistants 4 Admin Staff
- 41. Thanks to all the participants and their families NIHR UK Wellcome Trust MRC UK Cambridge and Peterborough NHS Foundation Trust