Introduction
There is an expanding interest in research with a specific focus on the experience of psychopathology in the realm of daily life. Developments in several scientific disciplines have fuelled the need for a momentary, daily life approach. In epidemiology, focus is shifting from observations of the macro-environment to the more detailed study of the micro-environment (McGrath, Reference McGrath2007). In psychology, there is an increased awareness that psychological models of psychopathology are essentially dynamic over time, requiring multiple assessments (Bentall, Reference Bentall2003). Cognitive science has focused on experiences as being embodied and situated, providing a powerful rationale for investigating experiences in the context in which they are occurring (Robbins & Aydede, Reference Robbins and Aydede2008). Finally, there is a growing need to ecologically validate experimental and laboratory findings (Koren et al. Reference Koren, Seidman, Goldsmith and Harvey2006). The study of persons in the context of normal daily life may provide a powerful and necessary addition to more conventional and cross-sectional research strategies in psychopathology (Table 1). The current paper provides a review of momentary assessment studies in psychopathology, with a focus on (1) phenomenology, (2) underlying mechanisms and aetiology, (3) dynamic psychological models, (4) biological mechanisms, (5) the application to treatment research and (6) gene–environment interactions.
Momentary assessment strategies
Momentary assessment strategies that are used most frequently are the Experience Sampling Method (ESM) (Delespaul, Reference Delespaul1995; Hektner et al. Reference Hektner, Schmidt and Csikszentmihalyi2007) and the Ecological Momentary Assessment (Stone & Shiffman, Reference Stone and Shiffman1994; Shiffman et al. Reference Shiffman, Stone and Hufford2008) (both methods hereafter called the ESM). The ESM is a structured diary method in which subjects are asked in normal daily life to report their thoughts, feelings and symptoms, and also the context (e.g. location, company, activity) and the appraisal of the context. Both open-ended questions and self-report Likert scales have been used. The reports typically have to be filled out several times a day (Fig. 1) during several consecutive days, either at random unpredictable moments, signalled by a beeper or alternatively, triggered by an event of interest (e.g. when someone is in contact with other people). In the paper-and-pencil version, subjects carry a digital wristwatch and a set of questionnaires collated in a booklet for each day. More recently, software packages have been developed for running these studies on palmtop computers (Schärer et al. Reference Schärer, Hartweg, Hoern, Graesslin, Strobl, Frey, Biedermann, Walser and Walden2002; Kimhy et al. Reference Kimhy, Delespaul, Corcoran, Ahn, Yale and Malaspina2006; Le et al. Reference Le, Choi and Beal2006; Granholm et al. Reference Granholm, Loh and Swendsen2008), providing equivalent results (Gwaltney et al. Reference Gwaltney, Shields and Shiffman2008).
Thought sampling is another quantitative method sampling daily life experiences, specifically quantifying characteristics of thinking (Hurlburt, Reference Hurlburt1990, Reference Hurlburt1997). Other ambulatory assessment technologies include instruments monitoring real-time biological phenomena (Fahrenberg & Myrtek, Reference Fahrenberg and Myrtek1996, Reference Fahrenberg and Myrtek2001) and end-of-day diaries, such as the ChronoRecord (Bauer et al. Reference Bauer, Wilson, Neuhaus, Sasse, Pfennig, Lewitzka, Grof, Glenn, Rasgon, Bschor and Whybrow2008), requiring one report a day for a number of days. As the current review aimed to investigate severe mental illness in terms of momentary mental processes in interaction with contextual features, only ESM studies are included that assess momentary self-reports at least twice a day.
Fig. 1. The Experience Sampling Method (ESM) provides multiple observations within one person, enabling researchers to investigate variation over time and patterns of covariation between different variables, e.g. variables y (––•––) and z (- -•- -).
Table 1. The pros and cons of the Experience Sampling Method (ESM) (see also Fig. 1)
Method
This review includes all ESM studies in psychopathology published since 1992, when the book The Experience of Psychopathology: Investigating Mental Disorders in their Natural Settings (de Vries, Reference de Vries1992) was published. The database PubMed was screened for published studies between 1992 and 2008 using the keywords ‘experience sampling method*’, ‘momentary assessment*’ and ‘ecological momentary assessment*’. Abstracts were screened and articles were selected that used the actual ecological momentary assessment or the ESM to study psychopathology with the exclusion of chronic pain and addiction. In addition, a first- and last-author bibliographical search was conducted for all papers resulting from the first search. Finally, several general ESM review papers were screened for papers on ESM and psychopathology (Moskowitz & Young, Reference Moskowitz and Young2006; Shiffman et al. Reference Shiffman, Stone and Hufford2008; www.ambulatory-assessment.org).
Phenomenology
The ESM enables a more detailed understanding of psychiatric phenomenology. When are psychiatric symptoms present? Do they fluctuate over time? How do they impact on daily functioning and what contextual conditions trigger, accompany or follow symptoms in the flow of daily life? This knowledge would allow for a better understanding of the phenomena, the underlying mechanisms, but also of the targets for treatment.
Affective disorders
Disturbed mood is the most prominent feature of mood disorders. However, there is little systematic information on how depression alters the dynamics of mood states in daily life. Peeters et al. (Reference Peeters, Berkhof, Delespaul, Rottenberg and Nicolson2006) found a more pronounced diurnal rhythm and more variability from moment to moment of negative affect in depressed patients compared to controls. Moreover, depressive persons exhibited increasing positive affect levels over the day with a later acrophase relative to healthy controls. This was replicated in a study investigating diurnal mood variation in depression in non-clinical individuals (Murray, Reference Murray2007). However, in this study the circadian function of positive affect was attenuated in the depressed group.
Depressed patients were found to experience lower levels of momentary quality of life that were mediated by momentary experiences of positive and negative affect, physical complaints and enjoyment of activities (Barge-Schaapveld et al. Reference Barge-Schaapveld, Nicolson, Berkhof and deVries1999). Depression also had an effect on work performance with significant decrements in task focus and productivity, equivalent to about 2.3 days of sickness per month (Wang et al. Reference Wang, Beck, Berglund, McKenas, Pronk, Simon and Kessler2004).
The ESM has also been used to study depression in children and adolescents (Axelson et al. Reference Axelson, Bertocci, Lewin, Trubnick, Birmaher, Williamson, Ryan and Dahl2003). One study reported more caffeine use and sleep problems in youth with major depression compared to healthy youngsters (Whalen et al. Reference Whalen, Silk, Semel, Forbes, Ryan, Axelson, Birmaher and Dahl2008). Youngsters with a depressive disposition were more likely to smoke, have the urge to smoke, and drink alcohol (Whalen et al. Reference Whalen, Jamner, Henker and Delfino2001). Depressive symptoms and problem behaviour have also been related to intense and labile emotions and less effective emotion regulation in adolescents (Silk et al. Reference Silk, Steinberg and Morris2003).
Only two studies used the ESM to investigate bipolar disorder. It was shown that patients with bipolar disorder spent more time in passive leisure activities and were more often alone at home compared to controls (Havermans et al. Reference Havermans, Nicolson and Devries2007). Mean levels of self-esteem and affect were not abnormal in remitted bipolar patients. However, bipolar patients showed strong fluctuations in both affect and self-esteem (Knowles et al. Reference Knowles, Tai, Jones, Highfield, Morriss and Bentall2007). The ESM thus allowed for a better understanding of dynamic mood changes over time and the situational determinants thereof.
Psychotic disorders
The phenomenology of positive symptoms of psychosis has been investigated extensively with the ESM. Delespaul et al. (Reference Delespaul, deVries and van Os2002) revealed that, contrary to what was expected, visual hallucinations in daily life were reported most frequently whereas auditory hallucinations were experienced with a higher intensity. The intensity of hallucinatory experiences was influenced by several contextual moderators, including social context and activity level (Delespaul et al. Reference Delespaul, deVries and van Os2002). Delusional moments were found to be present only one-third of the time and were accompanied by high negative and low positive affect. They were also found to be influenced by social factors and activity level (Myin-Germeys et al. Reference Myin-Germeys, Nicolson and Delespaul2001a). The importance of the dynamics of social context was emphasized by another ESM study (Verdoux et al. Reference Verdoux, Husky, Tournier, Sorbara and Swendsen2003b), showing that dynamic changes in social company were even stronger predictors of psychotic experiences in persons at risk for psychosis than social contact per se. These data thus show that psychosis is driven by subtle person–environment interactions in the stream of daily life.
Negative symptoms of psychotic disorders have also been studied using the ESM. Myin-Germeys et al. (Reference Myin-Germeys, Delespaul and deVries2000) studied flat affect in daily life and found schizophrenia patients to be more emotionally active and responsive than is assumed using behavioural and external observations. A study on anhedonia showed that schizophrenia patients displayed a deficit in anticipatory pleasure (pleasure related to future events) but not in consummatory pleasure (pleasure when directly engaging in an activity) (Gard et al. Reference Gard, Kring, Gard, Horan and Green2007). Furthermore, social anhedonia was characterized by increased time spent alone, a greater preference for solitude and less positive affect (Brown et al. Reference Brown, Silvia, Myin-Germeys and Kwapil2007). Psychosis-prone individuals were not more socially withdrawn, depressed or anxious in general. However, they spent more time doing nothing and reported higher levels of depression and anxiety in specific environmental or social situations (Husky et al. Reference Husky, Grondin and Swendsen2004). ESM data may thus provide a better insight into the true nature of the phenomena as they unfold in daily life.
Anxiety disorders
Only a few studies have been conducted in the area of anxiety disorders, a domain in which situational triggers are often of importance in eliciting symptoms. Dijkman-Caes et al. (Reference Dijkman-Caes, de Vries, Kraan and Volovics1993) compared behavioural aspects of panic disorder patients with and without agoraphobia and, contrary to expectation, found no difference between groups regarding the amount of time spent in public places. Panic subjects with agoraphobia did, however, spend significantly more time at home and with their families than did panic patients without agoraphobia and normal controls. Kenardy et al. (Reference Kenardy, Fried, Kraemer and Taylor1992) found that the prediction of panic was a significant precursor of an actual panic attack, but not the level of anxiety, threat or experience of control. Expectation of panic attacks was also associated with panic occurrence, as well as elevated sense of danger, helplessness, avoidance and catastrophic thoughts (Kenardy & Taylor, Reference Kenardy and Taylor1999). Expected panic attacks did not differ from unexpected attacks in the experience of the actual attack but, rather, levels of distress and arousal preceding the attack were higher in expected panic attacks.
De Beurs et al. (Reference de Beurs, Lange and Van Dyck1992) demonstrated that self-monitored incidence of panic attacks was lower than incidence rates obtained through retrospective estimation. Similarly, in patients diagnosed with obsessive–compulsive disorder (OCD), retrospective recall as opposed to ecological momentary assessment led to overestimation of symptom covariation with non-symptomatic variables such as stress and loneliness (Gloster et al. Reference Gloster, Richard, Himle, Koch, Anson, Lokers and Thornton2008). These findings challenge current diagnostic conceptualizations of panic disorder subtypes based on retrospective reports and call for more momentary research in the study of anxiety disorders and OCD.
Eating disorders
The feasibility and validity of using a momentary approach to test models of eating disordered behaviour have been demonstrated (Stein & Corte, Reference Stein and Corte2003; Engel et al. Reference Engel, Wonderlich, Crosby, Wright, Mitchell, Crow and Venegoni2005), emphasizing the importance of proximal antecedents (Smyth et al. Reference Smyth, Wonderlich, Crosby, Miltenberger, Mitchell and Rorty2001). Negative affect, and more specifically anger, in addition to low positive affect and dissociative experiences were found to precede a binge in the flow of daily life in both bulimia nervosa and binge eating disorder (Greeno et al. Reference Greeno, Wing and Shiffman2000; Alpers & Tuschen-Caffier, Reference Alpers and Tuschen-Caffier2001; Steiger et al. Reference Steiger, Gauvin, Engelberg, Ying Kin, Israel, Wonderlich and Richardson2005; Engel et al. Reference Engel, Boseck, Crosby, Wonderlich, Mitchell, Smyth, Miltenberger and Steiger2007; Smyth et al. Reference Smyth, Wonderlich, Heron, Sliwinski, Crosby, Mitchell and Engel2007; Stein et al. Reference Stein, Kenardy, Wiseman, Dounchis, Arnow and Wilfley2007; Wonderlich et al. Reference Wonderlich, Crosby, Engel, Mitchell, Smyth and Miltenberger2007a). Of note, several studies in patient samples and one in a student sample (Wegner et al. Reference Wegner, Smyth, Crosby, Wittrock, Wonderlich and Mitchell2002) found a deterioration of mood and self-perception after a binge, showing that binge eating may not be effective for regulating mood and self-perception (Steiger et al. Reference Steiger, Gauvin, Jabalpurwala, Seguin and Stotland1999, Reference Steiger, Gauvin, Engelberg, Ying Kin, Israel, Wonderlich and Richardson2005; Wegner et al. Reference Wegner, Smyth, Crosby, Wittrock, Wonderlich and Mitchell2002; Engelberg et al. Reference Engelberg, Steiger, Gauvin and Wonderlich2007; Hilbert & Tuschen-Caffier, Reference Hilbert and Tuschen-Caffier2007; Stein et al. Reference Stein, Kenardy, Wiseman, Dounchis, Arnow and Wilfley2007). Smyth et al. (Reference Smyth, Wonderlich, Heron, Sliwinski, Crosby, Mitchell and Engel2007), however, reported an increase in positive affect and a decrease in negative affect and anger after binge or vomit episodes in bulimia patients. A study that differentiated between the consequences of binging and purging behaviour reported a return to pre-binge mood state levels after purging but an increase in negative affect after binging. This study underlined that not the amount of negative mood but rather the association between negative mood and the desire to eat is specific for bulimia (Alpers & Tuschen-Caffier, Reference Alpers and Tuschen-Caffier2001). Binge episodes were found to be preceded by poorer than average self-concepts, poorer than average social experiences (Steiger et al. Reference Steiger, Gauvin, Engelberg, Ying Kin, Israel, Wonderlich and Richardson2005, Reference Steiger, Gauvin, Jabalpurwala, Seguin and Stotland1999) and potent family hassles (Okon et al. Reference Okon, Greene and Smith2003). Bulimic individuals also showed larger increases in self-criticism following negative social interactions, suggesting a hypersensitivity to social interactions in bulimia (Steiger et al. Reference Steiger, Gauvin, Jabalpurwala, Seguin and Stotland1999). Finally, dietary restraint has been shown to precede binge cravings whereas elevated binge cravings seem to predict actual binge episodes (Steiger et al. Reference Steiger, Gauvin, Jabalpurwala, Seguin and Stotland1999; Engelberg et al. Reference Engelberg, Gauvin and Steiger2005). This association, however, was not revealed in impulsive binge eaters (Steiger et al. Reference Steiger, Gauvin, Jabalpurwala, Seguin and Stotland1999).
Two studies used ESM data to validate different subtypes of bulimia nervosa by examining differences in disordered eating behaviour and patterns of affect. Wonderlich et al. (Reference Wonderlich, Crosby, Engel, Mitchell, Smyth and Miltenberger2007a) found increased levels of binging and purging on a daily basis in the interpersonal-emotional subtype, whereas Myers et al. (Reference Myers, Wonderlich, Crosby, Mitchell, Steffen, Smyth and Miltenberger2006) related multi-impulsive bulimia to increased impulsive and self-damaging behaviour but not to differences in eating behaviour compared to other subtypes of bulimia. It has also been shown that previous experience of childhood maltreatment may impact on mood and eating behaviours in bulimia nervosa (Wonderlich et al. Reference Wonderlich, Rosenfeldt, Crosby, Mitchell, Engel, Smyth and Miltenberger2007b). Sexual abuse was related to increased purging and self-destructive behaviours whereas emotional abuse was related more specifically to anger. It is important to understand this heterogeneity in bulimia nervosa in tailoring treatments to the styles and needs of different patients.
ESM studies on anorexia nervosa are less numerous. One case study (Vansteelandt et al. Reference Vansteelandt, Pieters, Vandereycken, Claes, Probst and Van Mechelen2004) revealed that a tendency to be physically active is related to emotional experiences and weight preoccupation in anorexia nervosa. A larger ESM study (Vansteelandt et al. Reference Vansteelandt, Rijmen, Pieters, Probst and Vanderlinden2007) indicated that lower body mass index was associated with higher levels of both the urge to be physically active and physical activity itself. In addition, self-discrepant beliefs may lead to affective lability in patients with anorexia nervosa, which in turn may lead to more restrictive behaviour and rituals. Both affect lability and restrictive behaviour were significantly exacerbated by stressful events (Engel et al. Reference Engel, Wonderlich, Crosby, Wright, Mitchell, Crow and Venegoni2005).
Attention deficit hyperactivity disorder (ADHD)
An ESM study on ADHD in adults revealed that the two ADHD symptom dimensions in adults were differentially associated with daily experiences (Knouse et al. Reference Knouse, Mitchell, Brown, Silvia, Kane, Myin-Germeys and Kwapil2008). Inattentive symptoms were positively associated with indices of general distress, whereas hyperactive-impulsive symptoms were unrelated to overall affective states and self-reported concentration problems but were associated with reduced sensitivity to contextual factors in perceptions of situations. Another study in children with ADHD and their mothers revealed that the preparatory tasks of daily living are especially challenging and linked disproportionately to child behaviour problems, parent negative affect and contentious interactions (Whalen et al. Reference Whalen, Jamner, Henker, Delfino and Lozano2002, Reference Whalen, Henker, Ishikawa, Jamner, Floro, Johnston and Swindle2006a). Mornings and weekends were found to be particularly challenging for ADHD children and their parents. Mothers of children with ADHD were found to have lower parenting esteem and elevated levels of anger in the company of their children (Whalen et al. Reference Whalen, Henker, Jamner, Ishikawa, Floro, Swindle, Perwien and Johnston2006b). A study investigating the effects of nicotine suggested that smokers with ADHD experience nicotine-related reductions in ADHD symptoms during daily life (Gehricke et al. Reference Gehricke, Whalen, Jamner, Wigal and Steinhoff2006).
Pervasive developmental disorder
Only one study used the ESM to investigate autism spectrum disorders. In 1994, the inner experiences of three persons with Asperger syndrome were investigated using the ESM. Two persons reported inner experiences whereas one was not able to report any inner experience. Thoughts were exclusively reported as images with almost no other features of inner experience (Hurlburt et al. Reference Hurlburt, Happe and Frith1994).
Borderline personality disorder (BPD)
Given the dynamic and reactive pattern of symptoms in BPD, the ESM is an attractive method for investigating its phenomenology. Borderline patients were found to report a higher frequency and increased intensity of negative affect (Stein, Reference Stein1996; Ebner-Priemer et al. Reference Ebner-Priemer, Kuo, Kleindienst, Welch, Reisch, Reinhard, Lieb, Linehan and Bohus2007; Russell et al. Reference Russell, Moskowitz, Zuroff, Sookman and Paris2007; Glaser et al. Reference Glaser, Van Os, Mengelers and Myin-Germeys2008) (Fig. 2), a lower frequency and decreased intensity of positive affect (Ebner-Priemer et al. Reference Ebner-Priemer, Kuo, Kleindienst, Welch, Reisch, Reinhard, Lieb, Linehan and Bohus2007; Glaser et al. Reference Glaser, Van Os, Mengelers and Myin-Germeys2008), increased levels of aversive tension (Stiglmayr et al. Reference Stiglmayr, Grathwol, Linehan, Ihorst, Fahrenberg and Bohus2005), and increased (Russell et al. Reference Russell, Moskowitz, Zuroff, Sookman and Paris2007; Reisch et al. Reference Reisch, Ebner-Priemer, Tschacher, Bohus and Linehan2008) and persistent (Ebner-Priemer et al. Reference Ebner-Priemer, Kuo, Schlotz, Kleindienst, Rosenthal, Detterer, Linehan and Bohus2008) anxiety and sadness in comparison to control subjects. In addition, increased levels of intra-individual variability and short-term fluctuations in overall affect valence (Stein, Reference Stein1996; Ebner-Priemer et al. Reference Ebner-Priemer, Kuo, Kleindienst, Welch, Reisch, Reinhard, Lieb, Linehan and Bohus2007; Russell et al. Reference Russell, Moskowitz, Zuroff, Sookman and Paris2007; Trull et al. Reference Trull, Solhan, Tragesser, Jahng, Wood, Piasecki and Watson2008), more instability on successive scores for hostility, fear and sadness (Trull et al. Reference Trull, Solhan, Tragesser, Jahng, Wood, Piasecki and Watson2008) and oscillation between these (Ebner-Priemer et al. Reference Ebner-Priemer, Kuo, Schlotz, Kleindienst, Rosenthal, Detterer, Linehan and Bohus2008; Reisch et al. Reference Reisch, Ebner-Priemer, Tschacher, Bohus and Linehan2008) were found to be important factors of the emotional dysregulation in BPD patients. Affective instability was also found to be associated with the experience of stress in the flow of daily life (Glaser et al. Reference Glaser, Van Os, Mengelers and Myin-Germeys2008), as were increased levels of dissociation (Stiglmayr et al. Reference Stiglmayr, Ebner-Priemer, Bretz, Behm, Mohse, Lammers, Anghelescu, Schmahl, Schlotz, Kleindienst and Bohus2008). BPD patients have, furthermore, been characterized by a pattern of negative recall (Ebner-Priemer et al. Reference Ebner-Priemer, Kuo, Welch, Thielgen, Witte, Bohus and Linehan2006). High levels of negative mood intensity were significantly related to impulsivity, self-reported suicidal ideation and suicidal behaviours over the past year in patients with BPD (Links et al. Reference Links, Eynan, Heisel, Barr, Korzekwa, McMain and Ball2007, Reference Links, Eynan, Heisel and Nisenbaum2008). Event-contingent recording research in BPD patients has revealed that their interpersonal behaviour is characterized by a tendency to jump between the pursuit and avoidance of interpersonal connectedness (Russell et al. Reference Russell, Moskowitz, Zuroff, Sookman and Paris2007). These results ecologically validate the incorporation of affective and interpersonal instability and stress reactive symptoms in the diagnosis of BPD.
Fig. 2. Real-life data from one patient with borderline personality disorder (––•––) and one control subject (- -•- -) retrieved from the study of Glaser et al. (Reference Glaser, Van Os, Mengelers and Myin-Germeys2008), showing differences between the two in level of negative affect and variability over time.
Phenomenology: conclusion
These data show that one of the most central characteristics of almost all symptoms observed in patients with severe psychiatric disorders is meaningful and widespread variation over time. Understanding this variation and the determinants thereof, both internal and situational, is therefore important for the diagnosis and treatment of these symptoms. Cross-sectional instruments are unable to assess this variation. As a consequence, data obtained with these instruments may not be linearly related to ESM findings.
Aetiology
Momentary assessment strategies may also constitute a powerful tool to examine underlying mechanisms relating to the onset and maintenance of psychopathology, especially when these mechanisms involve changes with regard to how persons react to or behave in certain situations or environments.
Stress
Stress has been postulated to play a role in all major psychiatric illnesses. Whereas previous research mainly focused on major life events, there is increasing evidence that the minor life events or daily hassles may be more powerful predictors of psychological symptoms (Monroe, Reference Monroe1983) and of subjective distress (Marco & Suls, Reference Marco and Suls1993). The ESM is a useful tool to study subjective appraisals of stressful events that occur in the context of daily life, in addition to the emotional and symptomatic reactions to these events, thus providing information on the stress–person interplay.
Stress sensitivity in daily life has been investigated extensively in relation to psychosis. It has been demonstrated that small stressors predict increased emotional reactions in persons with an increased liability to psychosis (Myin-Germeys et al. Reference Myin-Germeys, van Os, Schwartz, Stone and Delespaul2001b; Lataster et al. Reference Lataster, Wichers, Jacobs, Mengelers, Derom, Thiery, Van Os and Myin-Germeys2008), which was more pronounced in females (Myin-Germeys et al. Reference Myin-Germeys, Krabbendam, Delespaul and van Os2004). These subjects also show continuous variation in the intensity of subtle psychotic experiences associated with minor stresses in the flow of daily life (Myin-Germeys et al. Reference Myin-Germeys, Delespaul and van Os2005a). Furthermore, it has been demonstrated in two independent samples that altered stress sensitivity is independent of neurocognitive impairments, another well-known vulnerability marker for psychosis (Morrens et al. Reference Morrens, Krabbendam, Bak, Delespaul, Mengelers, Sabbe, Hulstijn, van Os and Myin-Germeys2007; Myin-Germeys et al. Reference Myin-Germeys, Krabbendam, Jolles, Delespaul and van Os2002). It is thus hypothesized that altered stress sensitivity is an independent ‘affective’ pathway to psychosis, possibly underlying the dimension of positive symptoms (Myin-Germeys & van Os, Reference Myin-Germeys and van Os2007). This affective pathway may partly result from a sensitization process by which previous exposures to severe stress, such as life events (Myin-Germeys et al. Reference Myin-Germeys, Krabbendam, Delespaul and van Os2003a), increase the sensitivity to small stresses in daily life, thus giving rise to a lasting behavioural liability (Collip et al. Reference Collip, Myin-Germeys and Van Os2008).
In affective disorders it was found that patients with current major depression experienced fewer positive events but the same amount of negative events compared to controls, although both events were rated as more stressful (Peeters et al. Reference Peeters, Nicolson, Berkhof, Delespaul and deVries2003b). Patients with remitted bipolar disorder, however, reported similar frequencies of daily hassles compared to controls and experienced these events as equally stressful. Higher levels of depression and more previous depressive episodes in the remitted bipolar patients predicted more stressful appraisals of daily events (Havermans et al. Reference Havermans, Nicolson and Devries2007). The results in the bipolar and depressed group were compared to similar data collected in patients with psychosis. Depressed patients showed a larger increase in negative affect, bipolar patients a larger decrease in positive affect, and psychotic patients both larger increases in negative and larger decreases in positive affect compared to controls (Myin-Germeys et al. Reference Myin-Germeys, Peeters, Havermans, Nicolson, DeVries, Delespaul and Van Os2003b). These findings show that emotional reactivity to daily stress may constitute part of the underlying vulnerability for severe mental illness, which is most pronounced in subjects suffering from psychosis.
Cannabis use
Cannabis use has been associated with psychosis. Verdoux et al. (Reference Verdoux, Gindre, Sorbara, Tournier and Swendsen2003a) found, in undergraduate students, that cannabis immediately increased the intensity of psychosis and decreased feelings of pleasure, especially in participants with high vulnerability for psychosis. Henquet et al. (unpublished observations) extended the finding that cannabis increases hallucinatory experiences to patients with psychosis. However, they found that patients are also more sensitive to the mood-enhancing effects of cannabis. Both studies found no evidence for the self-medication hypothesis because cannabis use was not predicted by previous mood or intensity of psychosis. Tournier et al. (Reference Tournier, Sorbara, Gindre, Swendsen and Verdoux2003) reported that cannabis use did not modify the level of state anxiety in daily life and, similarly, that the likelihood of using cannabis was not influenced by state anxiety. These findings provide a better insight into the motives for and the mechanisms of cannabis use, yielding essential information for the design and improvement of treatment strategies.
Aetiology: conclusion
Some of the mechanisms underlying psychopathology may be found in the dynamic interplay between the environment and the person. The longitudinal character of the data and the assessment of both experiences and environmental features enable the investigation of subtle person–environment interplays that subjects are not always consciously aware of.
Psychological models
Psychological models generally assume dynamic interactions between different constructs. However, most research is based on cross-sectional assessments. The longitudinal character of the ESM enables the study of dynamic patterns over time, thus capturing the film rather than a snapshot of daily life reality. Psychological modelling in daily life is still in its early stages but it seems a promising approach.
Psychological processes in paranoia
A recent psychological model states that persecutory delusions (e.g. the experience that someone wants to harm you) serve as a defence against low self-esteem, although studies investigating the relationship between paranoia and self-esteem level revealed inconsistent results (Bentall et al. Reference Bentall, Corcoran, Howard, Blackwood and Kinderman2001). A recent ESM study, however, examined the temporal relationship between paranoid experiences and both self-esteem level and variability in individuals with variable levels of paranoia (Thewissen et al. Reference Thewissen, Bentall, Lecomte, van Os and Myin-Germeys2008). Both lower self-esteem level and higher self-esteem variability were associated with higher levels of paranoia. Moreover, a temporal association between a decrease in momentary self-esteem and an increase in momentary paranoia was found. The use of the ESM, therefore, offers a plausible and intriguing explanation for the inconsistent results on the relationship between self-esteem level and paranoia, indicating that not self-esteem level per se but rather fluctuations in self-esteem may induce paranoia (Thewissen et al. Reference Thewissen, Bentall, Lecomte, van Os and Myin-Germeys2008).
Psychological processes in depression
The helplessness–hopelessness theory of anxiety and depression describes, in addition to the depressogenic attributional style (i.e. attributing negative events to stable and global causes), the cognitive aetiology of anxiety and the co-occurrence of both mood states. According to the theory, both anxiety and depression are characterized by helplessness (i.e. attributing that events are uncontrollable) but only depression is characterized by hopelessness (i.e. attributing that negative outcomes will be stable and global) (Alloy et al. Reference Alloy, Abramson, Metalsky and Hartlage1988). The ESM was used in analogue student samples to investigate the helplessness–hopelessness theory (Swendsen, Reference Swendsen1997, Reference Swendsen1998; Swendsen & Compagnone, Reference Swendsen and Compagnone2000). Attributional style in daily life did not explain mood fluctuations but was associated with the incidence of causal attributions and perception of controllability. Although the studies provided evidence that specific causal attributions of stability and globality about negative events explain fluctuations in depressed mood, there was no support for the relationship between uncontrollability attributions and anxious mood within the flow of daily life.
Coping
Coping is related to how people deal with the problems in their daily life (Lazarus & DeLongis, Reference Lazarus and DeLongis1983) and, as such, seems well suited for in-depth study with the ESM. However, relating the psychological concept of coping to real-life behaviour is complicated (Stone et al. Reference Stone, Schwartz, Neale, Shiffman, Marco, Hickcox, Paty, Porter and Cruise1998; Marco et al. Reference Marco, Neale, Schwartz, Shiffman and Stone1999). Only one study investigated ESM measures of coping in relation to psychopathology. In this study, in a sample of patients with psychotic disorder, symptomatic coping, defined as going along with the content of psychotic symptoms (e.g. carrying out hallucinatory commands), was compared to other, non-symptomatic forms of coping. It was shown that symptomatic coping, measured by semi-structured interview, was negatively associated with ESM measures of coping in daily life whereas non-symptomatic coping was positively associated with ESM measures of coping in daily life. In addition, it was suggested that effective coping may be associated with the tendency to develop conscious appraisals of distress associated with psychotic symptoms (Lardinois et al. Reference Lardinois, Myin-Germeys, Bak, Mengelers, van Os and Delespaul2007).
Treatment
Momentary assessment strategies may also have clear advantages in clinical psychopharmacology and treatment studies because they provide more accurate information retrieval and have superior ecological validity (Moskowitz & Young, Reference Moskowitz and Young2006). Furthermore, the collection of many data points reduces random error variance and increases sensitivity to detect change. In addition, daily diary methods may have added value for the assessment of medical adherence, especially in children and adolescents (Quittner et al. Reference Quittner, Modi, Lemanek, Ievers-Landis and Rapoff2008). Despite these clear advantages, there are few studies applying the ESM to investigate treatment effects in psychopathology. Barge-Schaapveld et al. (Reference Barge-Schaapveld, Nicolson, van der Hoop and De Vries1995) showed that clinical improvement was reflected in increased household activities, less passive leisure time and increased positive and decreased negative affect. After antidepressant treatment, responders, in contrast to non-responders, displayed a significant increase in activity level (Raoux et al. Reference Raoux, Benoit, Dantchev, Denise, Franc, Allilaire and Widlocher1994). In a study of imipramine (a tricyclic antidepressant agent) treatment on momentary quality of life, it was shown that the treatment condition did not change overall levels of quality of life but stabilized quality of life and decreased the level of inactivity (Barge-Schaapveld & Nicolson, Reference Barge-Schaapveld and Nicolson2002). Treatment also decreased sensitivity to stress and, more specifically, increased reward experience (Wichers et al. Reference Wichers, Barge-Schaapveld, Nicolson, Peeters, de Vries, Mengelers and van Os2008a). Response to treatment in depression may thus be conditional on restoration of hedonic capacity.
These studies show how the ESM may provide more fine-grained information with regard to treatment effects and adherence and more insight into the psychological mechanisms underlying response to treatment. Furthermore, studies performed in patients with generalized anxiety disorder and social phobia have shown computer-assisted ambulatory treatment to increase cost-effectiveness of cognitive behavioural therapy, while being efficient in reducing symptom levels (Newman et al. Reference Newman, Kenardy, Herman and Taylor1997, Reference Newman, Consoli and Taylor1999; Anderson et al. Reference Anderson, Jacobs and Rothbaum2004; Przeworski & Newman, Reference Przeworski and Newman2004).
Biological mechanisms
Momentary assessment strategies may also be useful for understanding variation in biological mechanisms as they play out in the flow of daily life. The current review focuses on papers that combine biological assessments with self-reports of subjective experiences in the flow of daily life in patients with severe mental illness.
Stress and the hypothalamic–pituitary–adrenal (HPA) axis
Stress responsivity is associated with HPA axis activity and the hormone cortisol (O'Connor et al. Reference O'Connor, O'Halloran and Shanahan2000) and dysfunction of the system has been associated with various psychopathologies (Watson & Mackin, Reference Watson and Mackin2006; Connan et al. Reference Connan, Lightman, Landau, Wheeler, Treasure and Campbell2007; Thompson et al. Reference Thompson, Phillips, Komesaroff, Yuen, Wood, Pantelis, Velakoulis, Yung and McGorry2007). Laboratory cortisol assessments fail to capture the complete dynamics of the cortisol pattern and, more importantly, cannot link the response to true ecological processes such as experiencing stress in the realm of daily life. As cortisol can be validly extracted from saliva (Kirschbaum & Hellhammer, Reference Kirschbaum and Hellhammer1993), it is well suited to be assessed using the ESM.
Thus far, only a small number of ESM studies have used the method of saliva sampling to investigate HPA axis activity in psychopathology (Peeters et al. Reference Peeters, Nicholson and Berkhof2003a, Reference Peeters, Nicolson and Berkhof2004; Lieb et al. Reference Lieb, Rexhausen, Kahl, Schweiger, Philipsen, Hellhammer and Bohus2004; Stetler et al. Reference Stetler, Dickerson and Miller2004; Stetler & Miller, Reference Stetler and Miller2005; Steptoe et al. Reference Steptoe, Gibson, Hamer and Wardle2007). In patients with major depressive disorder, the HPA axis response to negative daily events and mood changes was blunted (Peeters et al. Reference Peeters, Nicholson and Berkhof2003a). Erratic cortisol secretion was particularly present in out-patients with severe and recurrent forms of depression (Peeters et al. Reference Peeters, Nicolson and Berkhof2004). Stetler & Miller (Reference Stetler and Miller2005) found that a blunted cortisol response to awakening in depressed individuals might be due to a disruption of the regulatory influences of psychosocial factors and sleep patterns. Such daily activities, or ‘zeitgebers’, such as waking, eating or psychosocial factors were unrelated to cortisol secretion in depressed participants, suggesting a possible underlying mechanism of circadian disturbances in depression (Stetler et al. Reference Stetler, Dickerson and Miller2004). Furthermore, in a twin study, it was found that probands of co-twins with lifetime depression displayed a different diurnal cortisol profile than those without, suggesting that altered HPA axis functioning is an indicator of depression liability (Wichers et al. Reference Wichers, Myin-Germeys, Jacobs, Kenis, Derom, Vlietinck, Delespaul, Mengelers, Peeters, Nicolson and Van Os2008c). In patients diagnosed with BPD, ambulatory salivary cortisol measurements have suggested increased HPA axis activity, with higher total cortisol in response to awakening and higher total daily cortisol levels (Lieb et al. Reference Lieb, Rexhausen, Kahl, Schweiger, Philipsen, Hellhammer and Bohus2004).
Searching for biological markers
The ESM might be helpful to investigate biological markers for psychiatric disorders, by linking laboratory data to self-report assessments obtained in daily life.
Marcelis et al. (Reference Marcelis, Myin-Germeys, Suckling, Woodruff, Hofman, Bullmore, Delespaul and van Os2003) examined the relationship between cerebral tissue alterations and the experience of stress in the natural environment. Subjective stress experience was associated with increased levels of cerebrospinal fluid and reduced white-matter volumes in patients with psychosis (Marcelis et al. Reference Marcelis, Myin-Germeys, Suckling, Woodruff, Hofman, Bullmore, Delespaul and van Os2003). Another study in psychosis demonstrated that dopamine release induced by a physical stressor in the laboratory was associated with psychotic reactivity to daily life stress in first-degree relatives of patients with psychosis (Myin-Germeys et al. Reference Myin-Germeys, Marcelis, Krabbendam, Delespaul and van Os2005b). In depression, prefrontal cortex electroencephalography (PFC EEG) was combined with a momentary assessment study, showing that specific symptoms of depression such as rumination and self-esteem were uniquely associated with patterns of PFC EEG alpha activity (Putnam & McSweeney, Reference Putnam and McSweeney2008). In another study on depression, diminished late pupil dilation to negative words in a laboratory setting was found to be associated with higher levels of negative affect and lower levels of positive affect in the daily life of depressed children (Silk et al. Reference Silk, Dahl, Ryan, Forbes, Axelson, Birmaher and Siegle2007). Daily life research may thus drive the study of biological markers by improving our understanding of associated behavioural phenotypes.
Gene–environment interactions (G×E)
Finally, the ESM may be a useful tool in the rapidly developing field of the study of G×E in psychopathology, according to which susceptibility genes may amplify the risk for an individual to react with psychopathology in response to environmental pathogens. Moffitt et al. (Reference Moffitt, Caspi and Rutter2005) suggested a prospective collection of cumulative, repeated measures of proximal rather than distal environmental risk factors to optimize the environmental risk measurement in gene–environment research. ESM technology may be well placed to implement this methodological improvement.
Thus far, very few studies have used the ESM to study G×E in psychopathology. One study in psychosis investigated whether emotional reactivity to stress was mediated by the catechol-O-methyltransferase (COMT) Val158Met genotype. This study showed that patients with the Met/Met genotype showed the largest increase in psychotic symptoms and negative emotions in response to daily stressors, thus providing evidence for G×E mechanisms in the formation of psychotic symptoms (van Winkel et al. Reference van Winkel, Henquet, Rosa, Papiol, Faňanás, De Hert, Peuskens, van Os and Myin-Germeys2008). A second study in a general population sample found opposite results, with Val/Val carriers displaying more feelings of paranoia in response to stress (Simons et al. Reference Simons, Wichers, Derom, Thiery, Myin-Germeys, Krabbendam and van Os2008). A final study in patients with psychosis demonstrated that the COMT Val carriers, but not subjects with the Met/Met genotype, reported an increase in hallucinations in response to cannabis use (Henquet et al. Reference Henquet, Rosa, Delespaul, Papiol, Fananas, van Os and Myin-Germeys2008).
One other study used the ESM to study G×E in depression. In a general population twin study, it was demonstrated that probands whose co-twins were diagnosed with lifetime depression showed a stronger mood bias to stress than those with co-twins without such a diagnosis, thus providing evidence that the genetic liability to depression is in part expressed as the tendency to display negative affect in response to minor stressors in daily life (Wichers et al. Reference Wichers, Myin-Germeys, Jacobs, Peeters, Kenis, Derom, Vlietinck, Delespaul and van Os2007b). Of interest, however, the experience of positive affect associated with stress attenuated this endophenotypic expression of genetic vulnerability for depression (Wichers et al. Reference Wichers, Myin-Germeys, Jacobs, Peeters, Kenis, Derom, Vlietinck, Delespaul and van Os2007a). Similarly, positive affect decreased genetic moderation by a polymorphism in the gene encoding the brain-derived neurotrophic factor (BDNF; a factor known to be associated with stress-related behaviour) of the negative affect response to social stress (Wichers et al. Reference Wichers, Kenis, Jacobs, Myin-Germeys, Schruers, Mengelers, Delespaul, Derom, Vlietinck and van Os2008b).
These initial studies show the feasibility and power of ESM data for high-quality G×E studies examining the immediate effect of risk exposure conditional on genetic vulnerability for psychopathology in general as well as conditional on specific susceptibility polymorphisms.
Conclusion
This review has shown that variability over time and the dynamic interplay with the environment are essential features of psychopathological experiences that need to be captured for a better understanding of their phenomenology and underlying mechanisms. As variability is very difficult to assess with traditional instruments, the findings of ESM studies are a fundamental extension to cross-sectional data. The ESM has also been shown to contribute to a better understanding of the interplay between psychopathological experiences and environmental features, particularly because most of these interaction patterns are subtle and not consciously appraised.
Together, the data from the ESM literature suggest that the study of the variability and the dynamic patterns of reactivity of experiences, as they unfold in daily life, may accelerate our understanding of psychological and biological mechanisms underlying psychopathology. In addition, it may foster new research into treatment response and prove to be essential in the rapidly developing field of gene–environment interactions. The ESM enables capturing the film rather than the snapshot of daily life reality.
Declaration of Interest
None.
