Abstract
Background: Parental communication deviance (CD) has long been suggested as a potential risk factor for the development of psychosis and thought disorder in genetically sensitive offspring. However, the findings of the studies on the prevalence of CD in parents of psychotic patients have never been submitted to quantitative synthesis. Method: PsycINFO was searched from January 1959 to January 2012 for studies on the prevalence of CD in parents of psychotic patients. This search was supplemented with the results from a much larger systematic search (PsycINFO, PubMed, EMBASE, and Web of Science) on childhood trauma and psychosis. Results: A total of 20 retrieved studies (n = 1753 parents) yielded a pooled g of large magnitude (0.97; 95% CI [0.76; 1.18]) with a significant amount of heterogeneity (Q = 33.63; P = .014; I 2 = 46.47). Subgroup and sensitivity analysis of methodological features (study’s design, comparison group, diagnostic criteria, CD rating method, inter-rater reliability not reported, year of publication, and verbosity) and demographic characteristics (level of education or offspring’s age) revealed that pooled effect size was stable and unlikely to have been affected by these features. Conclusion: CD is highly prevalent in parents of psychotic offspring. This is discussed in the broader context of adoption and longitudinal studies that have reported a G × E interaction in the development of psychosis and thought disorder. A potential developmental mechanism is suggested to explain how CD may affect the developing offspring. The importance of further studies on CD and its potential value as a clinical concept are discussed.
Key words: communication deviance, thought disorder, family, psychosis
Daughter: What did you mean by a conversation having an outline? Has this conversation had an outline?
Father: Oh, surely, yes. But we cannot see it yet because the conversation isn’t finished. You cannot ever see it while you’re in the middle of it.
Bateson1 (p42).
Introduction
Over the last decade, there has been a renewed interest in the role of the environment in the development of psychotic experiences.2–5 It is becoming increasingly evident that a coherent scientific account of these experiences cannot be accomplished without the integration of environmental variables.6,7 Among these, quality of the family environment has long interested researchers.8–10 Family research in psychosis had its peak during the 1970s and 1980s with the publication of studies looking at different aspects of family interaction.11 Since then, interest in this field has somehow declined. This sociohistorical shift can be explained in part by 2 factors. The first factor is related to the emergence of neurobiological framework as a dominant paradigm of research in the field of psychosis.7 The second factor is related to sensitivities surrounding this line of research, and worries that this may lead to the stigmatization of families.12 It is our view that the family environment cannot be excluded as an important focus for both research and clinical intervention.13 Indeed, there is strong evidence that variables such as expressed emotion,14,15 family rearing environment,16,17 or family communication18,19 affect the course and development of psychosis.
One of the most researched family variables in the field of psychosis is parental communication deviance (CD).20 CD refers to a form of intrafamilial communication that is vague, fragmented, and contradictory and that compromises the development and sharing of meaning between the parent and the offspring, leading to the consequent breakdown in communication.21 The concept has a multidimensional structure22,23 and its frequency and severity are continuously distributed with no clear cutoff point.21,24 Examples can range from linguistic characteristics such as the use of contorted and peculiar language, eg, “This man is in the process in thinking of the process of being a doctor”25 (p166) and ambiguous linguistic referents, eg, “Kid stuff that’s one thing, but something else is different too” (D. I. Velligan, unpublished data, p. 18) to problems at the level of pragmatics such as non verbal disruptive behavior.26 Other areas of research have examined variables that can be assumed to be related to CD such as double-bind statements10 or thinking problems in the parents of psychotic offspring.27 However, some of these concepts have not been rigorously researched and, in some cases, the underlying concept does not necessarily reflect CD (eg, thought disorder).28,29
CD was initially developed and operationalized by Wynne and Singer,9,30–34 who devised a scoring system for the Rorschach and the Thematic Apperception Test.26 Since then, the field has evolved with the development of new methodologies,35,36 prospective cohort designs,18,37 and adoption studies that have helped elucidate the role of gene-environment interactions in determining the cross-generational transmission of psychotic communication disorders.19,38–42 The most important and replicated finding in the field is that CD is highly prevalent in the parents of psychotic offspring across diagnoses.24,25 Some authors have, therefore, suggested that exposure to this kind of communication during childhood may play a causal role in the development and ontogeny of psychotic experiences.21,43,44 However, there is still no known mechanism by which CD may affect the offspring. Some researchers have shown that CD can be successfully measured during problem solving where parent and psychotic offspring are asked to discuss a salient family problem.35,45,46 In other studies, it has been shown that parents with high scores on CD show less topic and affective focus during family discussions.47,48 Therefore, it is conceivable that the continuous exposure to communication that is vague, fragmented, and contradictory may lead the developing offspring to internalize it, resulting in psychotic experiences and thought disorder in particular.21,49 However, other mechanisms of transmission are clearly possible.
Since Wynne and Singer’s early work,30–34 the prevalence of CD among parents of psychotic offspring has been independently replicated by other groups using a variety of designs,18,19,28,36,50–55 with some studies documenting an exposure-response relationship between parental CD and severity of psychosis in the offspring.43,51,56,57 Other studies have shown more specific associations between CD and thought disorder,58 distractibility,59,60 and relapse.46,61 Moreover, CD does not seem to be a culture-bound phenomenon62–64 and has been found to be relatively uninfluenced by parents’ level of education65,66 or amount of speech produced (verbosity).28,43,67
Despite previous reviews,24,68–71 the literature in the field has never before been subjected to a quantitative analysis. Such analysis will allow us to establish the overall effect size for CD among parents of psychotic patients and determine its magnitude and consistency across studies. This analysis will also allow us to examine the impact of different study features that might affect our confidence in the findings. Finally, in conducting this analysis, we were also interested in comparing effect sizes between mothers and fathers. According to Wynne and Singer, the presence of at least 1 parent with low CD should be a protective factor and hence that CD is required in both parents for psychosis to develop. If this is the case, we would expect the magnitude of the effect to be the same for both parents. If, on the other hand, there is a difference in the magnitude of the effects for mothers and fathers, this might point to moderating factors that affect the mechanism by which CD confers risk of psychosis.
Methods
Literature Search Strategy and Eligibility Criteria
Two of the authors (F.V. and R.P.B.) undertook an initial search of all the published and unpublished materials on CD as part of a more extensive meta-analysis on childhood experiences and psychosis. Specific details of the search strategy can be found elsewhere.4 This initial search yielded 47 studies that were screened for coding at the phase 4 of the present search. To check this first search, which focused on a wide range of potential environmental risk factors, a second complementary search was undertaken specifically focusing on CD and psychosis. This was limited to the time period between January 1959 and January 2012. The starting year was chosen because this was the date of the first empirical study published on thought disturbance in parents of patients diagnosed with schizophrenia.27 PsycINFO was searched using the following search terms: “communication deviance,” “communication disturbance*,” and “thought dis*” combined with the terms famil*, parent*, mother, father combined with schiz*, and psycho* using Boolean operator “and” and “or.” To minimize publication bias, we included unpublished material in our search and tracked the citations of the most cited articles in the field.30,32–34 Concurrently, secondary searches were conducted using the names of the main publishing authors in the field, main methodologies, and research projects. Finally, we manually searched the bibliographic references of previous reviews in the field for material that had not been identified in our primary search24,68 and contacted available authors in the field for further information about their published and unpublished work.
One exclusion criterion was the use of “artificial family” designs.72–74 These designs involve experimental procedures in which family members of psychotic offspring interact with healthy offspring and parents of healthy offspring interact with psychotic offspring. These studies raised important issues about ecological validity and have used highly modified versions of Wynne and Singer’s methodology.
Also excluded were studies that operationalized concept formation using methodologies such as the object sorting test.27,75–80 In these studies, communication is not tested through means of a speech sample. Furthermore, they have already been subject to a meta-analysis and review.81 Studies that used analog measures of CD such as consensus tasks82–85 were also excluded. The dependent variable in these studies is the product of an interaction and so does not allow for a quantification of CD in parents individually. Another set of excluded studies were those where the dependent variable was disconfirmatory feedback or disqualifying communication.86–94 These constructs, despite measuring the quality of the family communication, only partly cover Wynne and Singer’s original construct. The final exclusion criterion involved studies that have measured subclinical thought disorder in parents of psychotic offspring28,95,96 because CD and thought disorder are not necessarily correlated in such samples.28
We initially retrieved a total of 22044 titles. Figure 1 shows the 4 phases of the systematic search. Only 20 studies were found to be eligible for analysis. In part, this was due to the fact that many of the studies retrieved were based on the same data set51,65,97,98 or were part of the same research project such as the UCLA High-Risk Project.18,47,48,99,100 In cases where studies had been based on the same data set, we selected the articles with the most complete statistical information for effect size computation.
Fig. 1.
Flowchart of studies included in meta-analysis.
Coding Protocol and Effect Sizes (g) Computation
The primary goal of the protocol was to allow a detailed subgroup analysis based on the methodological features that were likely to influence effect sizes. The following study characteristics were coded (see table 1): age of the offspring, diagnostic criteria, control group, type of methodology, and education or verbosity accounted for.
Table 1.
Characteristics of the Studies Included in the Meta-analysis
Source | Study Type | Parents (n) | Comparisons (n) | Control Groups | Education (Y/N) | Diagnostic Criteria (≥DSM-III) | Age of the Offspring (>15) | Methodology | Scoring | Parent | IRR (Y/N) | Verbosity (Y/N) |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Asarnow et al101 | CCS | 28 | 72 | Mixed | Y | Y | N | Projective | CD | Mother | Y | Y |
Behrens et al62 | CCS | 56 | 22 | Healthy | Y | N | Y | Projective | CD | Both | Y | N |
Docherty and Gordinier65 | CCS | 59 | 24 | Healthy | Y | Y | Y | Other | CDI | Both | Y | Y |
Docherty55 | CCS | 18 | 10 | Healthy | Y | Y | Y | Other | ICD | Both | Y | Y |
Glaser102 | CCS | 20 | 20 | Mixed | Y | N | Y | Projective | CD | Both | Y | Y |
Goldstein18 | PCS | 128 | N/a | Y | Y | Y | Projective | CD | Both | Y | Y | |
Hirsch and Leff103 | CCS | 40 | 40 | Other | Y | N | Y | Projective | CD | Both | Y | Y |
Holte and Wichstrøm93 | CCS | 14 | 28 | Mixed | Y | Y | Y | Projective | EU | Both | Y | Y |
Johnston and Holzman28 | CCS | 24 | 34 | Mixed | Y | N | Y | Projective | CD | Both | Y | Y |
Jones23 | CCS | 15 | 12 | Mixed | Y | N | Y | Projective | CD | Both | Y | Y |
Rund66 | CCS | 42 | 58 | Mixed | Y | Y | Y | Projective | CD | Both | Y | Y |
Sass et al58 | CCS | 42 | 8 | Other | Y | Y | Y | Projective | CD | Both | Y | Y |
Singer and Wynne31 | CCS | 40 | 80 | Other | N | N | N | Projective | CD | Both | N | N |
Solana67 | CCS | 40 | 40 | Other | Y | Y | Y | Projective | CD | Both | Y | Y |
Wender et al104 | CCS | 56 | 28 | Other | N | N | Y | Projective | CD | Both | Y | Y |
Wild et al52 | CCS | 44 | 49 | Healthy | Y | N | Y | Other | CD | Both | N | N |
Wild et al50 | CCS | 72 | 102 | Mixed | Y | N | Y | Other | CD | Both | Y | Y |
Wynne105 | CCS | 38 | 80 | Mixed | N | N | Y | Projective | CD | Both | N | N |
Wynne et al43 | CCS | 88 | 140 | Mixed | Y | N | Y | Projective | CD | Both | Y | Y |
Singer et al22 | CCS | 26 | 16 | Other | N/K | N | Y | Projective | CD | Both | N | Y |
Note: CCS, case-control study; CD, communication deviance; CDI, Cognitive Disturbances Index; DSM-III, Diagnostic and Statistical Manual of Mental Disorders-III; EU, egocentric utterances; healthy, parents of healthy offspring; ICD, interactional communication deviance; IRR, inter-rater reliability reported; mixed, multiple contrast groups; N, no; N/a, not applicable; N/K, not known; other, parents of patients with other mental health diagnoses; PCS, prospective cohort study; projective, Rorschach, TAT, or Phillipson; Y, yes.
The computation of effect sizes and consequent statistical analysis were performed using Comprehensive Meta-analysis.106 This software allows the user to easily compute effect sizes using a wide variety of data formats.
The computation of effect sizes was performed using hedges’ g 107 given that CD is a continuous construct with no real threshold values. In some studies, the means, SDs, samples sizes, and P values for both the experimental and control groups were available43,51 and therefore g was calculated using the original equation.
![]() |
Here, Me stands for mean of the experimental group, Mc for mean of the control group, and N for number of participants. In 2 cases,22,58 SDs were not available and g was calculated using the means, sample sizes, and P values for both groups of parents. In 1 case,67 g was computed from t test, P value, and effect direction given that no other statistical data were available.
In studies where the dependent variable was dichotomous,18,23,31,50,52,62,101,104 standard means difference (SMD) was calculated from OR using agreed statistical conventions.108,109
![]() |
Finally, in one study, g was calculated from the chi-square (χ2), P value, and sample size105 given that no other statistical information was available. In studies where the research design included more than 2 groups, the effect size was calculated from the comparison between parents of offspring diagnosed with schizophrenia and parents of healthy offspring.28,43,66 For studies that had used more than 1 concept of CD, only the data that reflected Wynne and Singer’s original conceptualization were extracted.
Finally, computation of effect sizes was carried out under random effects model given that our assumption was that the studies retrieved were likely to be heterogeneous and the analysis was likely to carry across-study variation.110
Results
Pooled Effect Sizes and Heterogeneity Analysis
The final analysis included 19 case-control studies and data from 1 prospective study.18 The studies included an overall pooled sample of 1753 parents. The computation carried out under a random effects model for the entire sample revealed a very large pooled effect size of 1.45 (SE = 0.27; 95% CI [0.92; 1.97]; z = 5.41; P < .001), with a significant amount of heterogeneity (Q[19] = 238.8; P < .001; I 2 = 92.04). After visual inspection of the funnel plot, we decided to exclude Wynne et al43 because this effect size was of an unusually large magnitude (g = 12.4; SE = 0.8; 95% CI [10.84; 13.97]; z = 15.54; P < .001). The exclusion of this outlier reduced the overall effect size to 0.97 (SE = 0.11; 95% CI [0.76; 1.18]; z = 9.2; P < .001), with a more acceptable, but still significant, level of heterogeneity (Q[18] = 33.63; P = .014; I 2 = 46.47; τ2 = 0.1; τ = 0.3). According to benchmark thresholds,111 we can interpret that our pooled g is of large magnitude. One-study removed analysis revealed that the results were stable and unlikely to be affected by the exclusion of any one study (figure 2).
Fig. 2.
Forest plot.
In order to test how the different features affected our result, we recomputed the pooled effect size extracting studies using our coding protocol. The exclusion of 1 cohort study did not change our overall effect size (k = 18; g = 0.96; SE = 0.11; 95% CI [0.75; 1.16]; z = 8.95; P < .001; Q[17] = 32.42; P = .013; I 2 = 47.56; τ2 = 0.09; τ = 0.3). The computation of the effect size using studies that had compared parents of psychotic offspring with healthy controls (as opposed to other kinds of controls, eg, parents of children with depression or learning difficulties), again, did not change the pooled effect size significantly (k = 5; g = 0.91; SE = 0.24; 95% CI [0.44; 1.38]; z = 3.79; P < .001; Q[4] = 14.76; P = .005; I 2 = 72.91; τ2 = 0.2; τ = 0.45). However, heterogeneity increased significantly and the CI broadened. The exclusion of studies that had not controlled for the parents’ educational level brought the effect size to a still significant and large 0.89 (k = 15; SE = 0.11; 95% CI [0.67; 1.1]; z = 7.97; P < .001; Q[14] = 22.29; P = .073; I 2 = 37.18; τ2 = 0.06; τ = 0.25). The exclusion of studies that had tested parents of children below the age of 15 again did not change the overall pooled effect size (k = 17; g = 0.94; SE = 0.1; 95% CI [0.73; 1.14]; z = 8.99; P < .001; Q[16] = 27.39; P = .037; I 2 = 41.59; τ2 = 0.07; τ = 0.27). Unfortunately, we could not carry out a subgroup analysis by offspring’s gender because there were too few data. Finally, we recomputed the effect size for studies that had accounted for verbosity. This reanalysis brought the pooled g down to a still large and significant 0.83 (k = 15; SE = 0.11; 95% CI [0.63; 1.04]; z = 7.93; P < .001; Q[14] = 20.09; P = .127; I 2 = 30.32; τ2 = 0.05; τ = 0.22), whereas the effect size for studies that did not account for verbosity was 1.35 (k = 4; SE = 0.16; 95% CI [1.03; 1.66]; z = 8.42; P < .001; Q[3] = 3.31; P = .346; I 2 = 9.45; τ2 = 0.01; τ = 0.1) and difference between the 2 types of study was significant (Q[1] = 7.22; P = .007).
Because diagnostic criteria have changed considerably since Wynne and Singer’s early studies, we decided to run a subgroup analysis of the studies carried out before and after the publication of Diagnostic and Statistical Manual of Mental Disorders-III (DSM-III). The estimated effect size for the latter studies was 0.96 (k = 8; SE = 0.2; 95% CI [0.57; 1.36]; z = 4.78; P < .001), however, with a significant level of heterogeneity (Q[7] = 16.75; P = .019; I 2 = 58.22; τ2 = 0.18, τ = 0.42). The pooled effect size for the studies undertaken before the publication of the DSM-III was 0.98 (k = 11; SE = 0.12; 95% CI [0.74; 1.22]; z = 8.02; P < .001), but with a nonsignificant level of heterogeneity (Q[10] = 16.37; P = .09; I 2 = 38.9). To complement the above analysis, we carried out a metaregression using year of publication as a moderator variable. The regression was carried out using mixed effects to allow for between-study heterogeneity. Overall, year of publication was not found to be a significant predictor of effect size (β = −0.01; SE = 0.01; 95% CI [−0.03; 0.02]; z = −0.51; P = .61; α = 11.68; SE = 20.99; 95% CI [−29.47; 52.82]; z = 0.56; P = .58; τ2 = 0.1; Q R = 0.26; P = .61; Q E = 14.82; P = .61; Q T = 15.08; P = .66).
Another potential source of heterogeneity between studies was the diverse methodologies used to elicit speech samples. A subgroup analysis of the studies that have used projective techniques revealed a g = 0.93 (k = 16; SE = 0.13; 95% CI [0.68; 1.18]; z = 7.17; P < .001) with a significant amount of heterogeneity (Q[15] = 28.49; P = .012; I 2 = 50.87; τ2 = 0.12; τ = 0.34), whereas studies that have used other methodologies yielded a g of 1.09 (k = 4; SE = 0.15; 95% CI [0.79; 1.39]; z = 7.05; P < .001) with a nonsignificant level heterogeneity (Q[3] = 3.45; P = .327; I 2 = 13.04).
We were also interested in the effect sizes per parental sex. The effect size for mothers of psychotic offspring was g = 0.89 (k = 7; SE = 0.18; 95% CI [0.54; 1.24]; z = 4.99; P < .001; Q[6] = 7.92; P = .244; I 2 = 24.21). Analysis of the effect sizes for the fathers of psychotic offspring revealed a much smaller g = 0.39 (k = 6; SE = 0.16; 95% CI [0.07; 0.7]; z = 2.42; P < .05; Q[5] = 5.06; P = .41; I 2 = 1.2). Using a mixed effect analysis, the comparison revealed that the difference between the 2 mean effect sizes was statistically significant (Q[1] = 4.38; P < .05).
Finally, we ran subgroup analysis between studies that have reported inter-rater reliability (IRR) (k = 14; g = 0.87; SE = 0.12; 95% CI [0.64; 1.11]; z = 7.25; P < .001; Q[13] = 21.49; P = .064; I 2 = 39.5; τ2 = 0.07; τ = 0.27) and studies that either had not reported IRR or in which reliability was poor104 (k = 5; g = 1.18; SE = 0.17; 95% CI [0.84; 1.52]; z = 6.83; P < .001; Q[4] = 6.5; P = .165; I 2 = 38.45; τ2 = 0.06; τ = 0.24). Using a mixed effect analysis, the comparison revealed that the difference between the 2 mean effect sizes was not statistically significant (Q[1] = 2.1; P = .147).
Publication Bias
The visual inspection of the funnel plot revealed potential publication bias. The following statistical tools were used to explore this: (1) Begg and Mazumdar’s rank order correlation, (2) Egger’s regression intercept, and (3) Duval and Tweedie’s “trim and fill” procedure. Begg and Mazumdar’s test112 revealed a nonsignificant Kendall’s τ of 0.22 (z = 1.33; P = .09). Although this nonsignificant value suggests nonexistence of publication bias, the test has relatively low power for meta-analyses with a small number of studies.
In Egger’s test,113 funnel plot asymmetry was calculated from a linear regression where the more the intercept deviates from 0, the more pronounced the asymmetry. Given the small number of pooled studies, evidence for asymmetry was taken at P < .1. In our case, the intercept was 1.49 (95% CI [−0.54; 3.52]; t[17] = 1.55; P = .07) supporting the existence of bias. Finally, Duval and Tweedie’s114 “trim and fill” procedure identified 2 potential missing studies. The recomputed point estimate was 0.91 (95% CI [0.7; 1.13]) revealing that the presence of these “missing studies” was not likely to affect the overall magnitude of our effect size.
Discussion
The overall pooled effect size of the studies we considered was large in magnitude supporting Wynne and Singer’s clinical intuitions and early findings. This effect was observed to be larger for mothers than for fathers.
It can be argued that these results should be treated cautiously given that there was considerable heterogeneity between studies. However, we have tried to circumvent this problem by combining individual effect sizes using a random effects model by excluding an outlier and finally through subgroup and sensitivity analyses, which brought down the heterogeneity to more acceptable levels. The most striking consequence was that all of these analyses continued to reveal large and significant effect sizes suggesting that variation between studies should not undermine our overall confidence in the association between CD in parents and psychosis in offspring. It is important to note, however, that the result of Egger’s test suggested the existence of publication bias. When we recomputed a point estimate using the “trim and fill” procedure, we found that potential missing studies were not likely to undermine the presence of a significant effect. However, the existence of publication bias is an important issue that cannot be disregarded completely.
Some between-study variability should be unsurprising given the many limitations that these studies have in terms of their methodological quality. For example, many of the studies that were undertaken before the publication of DSM-III used diagnostic approaches that were quite different from today’s standardized methods. On this issue, it was reassuring to see that a subgroup analysis, based on this specific methodological variable, yielded stable effect sizes across groups. Also, metaregression analysis showed that year of publication of the study had no impact on overall effect size.
A second problem affecting some of the old studies is the multiplicity of different methodologies used to gather speech samples. This plurality has the advantage of showing that effect sizes are not an artifact of any specific methodology. However, it makes the task of quantitative synthesis difficult and is likely to be a source of significant amount of heterogeneity. This is especially relevant when one looks at the differences in the methods of standardization of the CD measures. For example, in some studies, this ratio was calculated from word counts,103 whereas in others studies, it was calculated using the number of lines of speech.55 On this specific limitation, it was reassuring to note that the magnitude of the effect size did not change when we subanalyzed the results using methodology as a criterion. It is also interesting that some studies have replicated Wynne and Singer’s early findings using sophisticated and standardized linguistic methodologies51 showing that findings on CD can be replicated using more natural conversational samples.
Another issue that limited the power of the present meta-analysis is the lack of standardized threshold values. In the studies in which results were reported as dichotomous outcomes, we have opted to reconstruct a continuous variable using agreed statistical conventions. However, we are aware that this approach has some limitations. In this respect, we felt reassured by the fact that nearly all the studies identified and retrieved reported positive findings and by the fact that our analysis comparing dichotomous and continuous outcomes revealed a stable effect size.
Another limitation is related to issues of reliability and validity. If it is true that in the majority of studies IRR has been ascertained and reported, it is no less true that in a few studies this issue, if not completely ignored, at least was not discussed and reported. Our subgroup analysis revealed that differences in mean effect sizes between studies that reported IRR and studies that did not were not significant.
Finally, our analysis tells us very little about the explanation for the prevalence of CD in parents of psychotic offspring. Given the robustness of the association between psychosis in offspring and parental CD, it is pertinent to consider the possible processes that might account for it.
Speculating about reverse causality, some authors have suggested that the higher prevalence of CD in parents of psychotic offspring could reflect a reaction of the parent to the disordered communication of the offspring.69,72–74,115,116 However, rigorous studies have shown that parents’ CD measured during individual protocols correlates positively with CD during problem-solving situations with their offspring,35 suggesting that variance in CD is not explained by the offspring’s immediate behavior.40,102
In an attempt to settle the issue, prospective and adoption studies have demonstrated that CD in the parent precedes the development of psychosis in the offspring by many years,18,19 and more importantly, that healthy communication in an adopting couple seems to exert a protective effect in the case of high-risk adoptees.41 Also relevant to the question of reverse causality is the observation that CD has the quality of an enduring trait-like characteristic in parents45,61,117 that becomes stable in the transition from adolescence to adulthood118 and that does not worsen with arousal or stress in the parent.56,98 These studies as a whole support the view that CD, rather than a reactive and transient phenomenon,80 seems to be an enduring characteristic of some parents.61,117 Despite these observations, it remains possible that CD may become involved in a complex dynamic process of circular causality where cause and effect are intertwined12,94 as appears to be the case with other family variables such as expressed emotion.119–121
Another hypothesis that has been suggested to explain the prevalence of CD among parents of psychotic offspring is that this form of communication could be an epiphenomenon of shared genetic vulnerability to psychosis, ie, endophenotype.104,122,123 According to this account, CD among parents of psychotic offspring should be interpreted against a broader context of cognitive deficits that are believed to be an expression of genetic liability for schizophrenia among unaffected first-degree relatives.124
Some researchers have suggested that the FOXP2 gene (CNTNAP2 pathway) could be responsible for a shared vulnerability to CD and thought disorder125,126 although a recent meta-analysis of 2 genome-wide association studies found a significant association between thought disorder and 4 other genetic loci (PKNOX2, MYH13, PHF2, and GPC6).127 However, an exclusively genetic account seems unlikely given that methodologically rigorous studies have shown that CD is a transdiagnostic risk factor for psychiatric disorders other than schizophrenia.128,129 Furthermore, in our meta-analysis, a larger effect was found for maternal CD than for paternal CD. Although sex-linked genetic effects are not impossible, an obvious explanation for this finding is that actual exposure to CD is required for there to be an increased risk in the offspring. This gender effect appears to suggest that low CD in one of the parents (fathers) by itself may not have the protective impact on the development of psychosis that Wynne and Singer initially hypothesized. However, we were not able to extract data comparing families in which both parents had CD vs families in which only 1 parent had CD, and we therefore could not test the protection hypothesis directly.
Interestingly, evidence of a gene-environment interaction was reported by Tienari and colleagues in a relatively small study,19,41,42 in which it was found that at-risk children only developed a psychosis and thought disorder if they were raised by adoptive parents who exhibited CD. This finding has never been replicated and, if found to be secure in further studies, would provide one of the few examples of gene-environment interactions known to be important in psychosis. Hence, an important avenue for future research might be to measure parental CD alongside genetic data obtained from offspring in studies comparing patients and controls.
Prospective cohort studies could help us clarify the mechanism by which exposure to CD may affect the development of psychosis19 and thought disorder41,42 in offspring, which might ultimately have important implications for the prevention of psychosis and thought disorder. Miklowitz and Stackman24 have suggested that exposure to CD might act as a psychosocial stressor that particularly affects genetically sensitive children. If this is the case, it might be expected that CD will affect the likelihood of a wide range of psychotic and affective symptoms. An alternative possibility is that CD affects some symptoms more than others. Indeed, some studies have focused specifically on the relationship between parental CD and thought disorder in offspring58 although, to our knowledge, none have carried out adequate statistical controls for the co-occurrence of symptoms. Since Wynne and Singer first began their work on CD, more has been discovered about the structure of psychosis and its course across time. Although more complex models have also been proposed,130 a widely supported model proposes that all psychotic disorders can be described along 5 symptom dimensions: positive symptoms, negative symptoms, cognitive disorganization, depression, and mania.131 Future research should consider whether CD is specifically related to any of these dimensions.
Consistent with the idea that actual exposure to CD is required to increase the risk of psychosis, some authors have argued that the impact of maternal CD is likely to occur during early development through the progressive internalization of language during social interaction with the parent,21,29,49,132 a hypothesis that is consistent with Vygotsky’s sociocultural analysis of cognitive development.133 Another hypothesis is that the inability to establish and maintain shared foci of attention in the parent has a specific effect on very early nonverbal reciprocal dialogues between mother and baby, leading to the disruption of early cognitive development and communication in the child, resulting in a high risk of psychosis.134 Interestingly, and despite remaining completely unexplored, this hypothesis is consistent with work by researchers in the field of developmental psychology, whose studies have demonstrated the crucial role of the mother in scaffolding the child’s cognitive development during early episodes of joint attention.135–143 Along with other ante- and perinatal risk factors that have been suggested within the framework of the neurodevelopmental hypothesis of schizophrenia, such a developmental pathway could in part explain the results of birth cohort studies that have documented early developmental delays in children who were later diagnosed with schizophrenia.144–146 This hypothesis is also consistent with studies that show an association between maternal CD and poor cognition in offspring diagnosed with schizophrenia49,59,60 and with our present finding of a lower prevalence of CD in fathers of psychotic offspring. This hypothesis should be considered in future cohort studies.
At a clinical level, the revival of research on CD may have important implications for the development of family interventions. Results from a systematic review147 and meta-analyses148,149 have clearly shown that these interventions are effective in reducing psychotic relapse and hospital admissions. Parental CD has also been found to be associated with psychotic relapse in the offspring in 2 independent studies.46,61 Hence, a fruitful avenue for enhancing these interventions might be to target the quality of the family communication from a CD standpoint.
Funding
Fundação para a Ciência e a Tecnologia (SFRH/BD/ 77379/2011 to P.S.).
Acknowledgment
The authors have declared that there are no conflicts of interest in relation to the subject of this article.
References
- 1. Bateson G. Steps to an Ecology of Mind: Collected Essays in Anthropology, Psychiatry, Evolution, and Epistemology. San Francisco, CA: Chandler Publishing Company; 1972. [Google Scholar]
- 2. van Os J, Krabbendam L, Myin-Germeys I, Delespaul P. The schizophrenia envirome. Curr Opin Psychiatry. 2005;18:141–145 [DOI] [PubMed] [Google Scholar]
- 3. Bentall RP. Madness Explained: Psychosis and Human Nature. London, UK: Allen Lane; 2003. [Google Scholar]
- 4. Varese F, Smeets F, Drukker M, et al. Childhood adversities increase the risk of psychosis: a meta-analysis of patient-control, prospective- and cross-sectional cohort studies. Schizophr Bull. 2012;38:661–671 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. van Os J, Kenis G, Rutten BPF. The environment and schizophrenia. Nature. 2010;468:203–212 [DOI] [PubMed] [Google Scholar]
- 6. Bentall RP, Fernyhough C. Social predictors of psychotic experiences: specificity and psychological mechanisms. Schizophr Bull. 2008;34:1012–1020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Read J, Bentall RP, Fosse R. Time to abandon the bio-bio-bio model of psychosis: exploring the epigenetic and psychological mechanisms by which adverse life events lead to psychotic symptoms. Epidemiol Psychiatr Soc. 2009;18:299–310 [PubMed] [Google Scholar]
- 8. Lidz T, Cornelison A, Terry D, Fleck S. Intrafamilial environment of the schizophrenic patient. VI. The transmission of irrationality. AMA Arch Neurol Psychiatry. 1958;79:305–316 [DOI] [PubMed] [Google Scholar]
- 9. Wynne LC, Ryckoff IM, Day J, Hirsch SI. Pseudo-mutuality in the family relations of schizophrenics. Psychiatry. 1958;21:205–220 [DOI] [PubMed] [Google Scholar]
- 10. Bateson G, Jackson DD, Haley J, Weakland J. Toward a theory of schizophrenia. Behav Sci. 1956;1:251–264 [Google Scholar]
- 11. Doane JA. Family interaction and communication deviance in disturbed and normal families: a review of research. Fam Process. 1978;17:357–376 [DOI] [PubMed] [Google Scholar]
- 12. Read J, Seymour F, Mosher LR. Unhappy families. In: Read J, Mosher LR, Bentall RP, eds. Models of Madness: Psychological, Social and Biological Approaches to Schizophrenia. New York, NY: Brunner-Routledge; 2004:253–268 [Google Scholar]
- 13. Lobban F, Barrowclough C. A Casebook of Family Interv entions for Psychosis. Oxford, UK: Wiley-Blackwell; 2009. [Google Scholar]
- 14. Butzlaff RL, Hooley JM. Expressed emotion and psychiatric relapse: a meta-analysis. Arch Gen Psychiatry. 1998;55:547–552 [DOI] [PubMed] [Google Scholar]
- 15. Hooley JM. Expressed emotion and relapse of psychopathology. Annu Rev Clin Psychol. 2007;3:329–352 [DOI] [PubMed] [Google Scholar]
- 16. Carter JW, Schulsinger F, Parnas J, Cannon T, Mednick SA. A multivariate prediction model of schizophrenia. Schizophr Bull. 2002;28:649–682 [DOI] [PubMed] [Google Scholar]
- 17. Tienari P, Wynne LC, Sorri A, et al. Genotype-environment interaction in schizophrenia-spectrum disorder. Long-term follow-up study of Finnish adoptees. Br J Psychiatry. 2004;184:216–222 [DOI] [PubMed] [Google Scholar]
- 18. Goldstein M. The UCLA high-risk study. Schizophr Bull. 1987;13:505–514 [DOI] [PubMed] [Google Scholar]
- 19. Wahlberg KE, Wynne LC, Hakko H, et al. Interaction of genetic risk and adoptive parent communication deviance: longitudinal prediction of adoptee psychiatric disorders. Psychol Med. 2004;34:1531–1541 [DOI] [PubMed] [Google Scholar]
- 20. Tienari P, Wahlberg KE. Family environment and psychosis. In: Morgan C, McKenzie K, Fearon P, eds. Society and Psychosis. Cambridge, UK: Cambridge University Press; 2008:112–126 [Google Scholar]
- 21. Wynne LC. Current concepts about schizophrenics and family relationships. J Nerv Ment Dis. 1981;169:82–89 [DOI] [PubMed] [Google Scholar]
- 22. Singer MT, Wynne LC, Toohey ML. Communication disorders and the families of schizophrenics. In: Wynne LC, Cromwell RL, Matthysse S, eds. The Nature of Schizophrenia: New Approaches to Research and Treatment. New York, NY: John Wiley & Sons; 1978:499–511 [Google Scholar]
- 23. Jones JE. Patterns of transactional style deviance in the TAT’s of parents of schizophrenics. Fam Process. 1977;16:327–337 [DOI] [PubMed] [Google Scholar]
- 24. Miklowitz DJ, Stackman D. Communication deviance in families of schizophrenic and other psychiatric patients: current state of the construct. Prog Exp Pers Psychopathol Res. 1992;15:1–46 [PubMed] [Google Scholar]
- 25. Miklowitz DJ, Velligan DI, Goldstein MJ, et al. Communication deviance in families of schizophrenic and manic patients. J Abnorm Psychol. 1991;100:163–173 [DOI] [PubMed] [Google Scholar]
- 26. Singer MT, Wynne LC. Principles for scoring communication defects and deviances in parents of schizophrenics: Rorschach and TAT scoring manuals. Psychiatry. 1966;29:260–288 [PubMed] [Google Scholar]
- 27. McConaghy N. The use of an object sorting test in elucidating the hereditary factor in schizophrenia. J Neurol Neurosurg Psychiatry. 1959;22:243–246 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Johnston MH, Holzman PS. Assessing Schizophrenic Thinking: A Clinical and Research Instrument for Measuring Thought Disorder. 1st ed. San Francisco, CA: Jossey-Bass; 1979. [Google Scholar]
- 29. Wynne LC. The epigenesis of relational systems: a model for understanding family development. Fam Process. 1984;23:297–318 [DOI] [PubMed] [Google Scholar]
- 30. Wynne LC, Singer MT. Thought disorder and family relations of schizophrenics. II. A classification of forms thinking. Arch Gen Psychiatry. 1963;9:199–206 [DOI] [PubMed] [Google Scholar]
- 31. Singer MT, Wynne LC. Differentiating characteristics of parents of childhood schizophrenics, childhood neurotics, and young adult schizophrenics. Am J Psychiatry. 1963;120:234–243 [DOI] [PubMed] [Google Scholar]
- 32. Singer MT, Wynne LC. Thought disorder and family relations of schizophrenics: III. Methodology using projective techniques. Arch Gen Psychiatry. 1965;12:187–200 [DOI] [PubMed] [Google Scholar]
- 33. Singer MT, Wynne LC. Thought disorder and family relations of schizophrenics. IV. Results and implications. Arch Gen Psychiatry. 1965;12:201–212 [DOI] [PubMed] [Google Scholar]
- 34. Wynne LC, Singer MT. Thought disorder and family relations of schizophrenics: I. A research strategy. Arch Gen Psychiatry. 1963;9:191–198 [DOI] [PubMed] [Google Scholar]
- 35. Velligan DI, Goldstein MJ, Nuechterlein KH, Miklowitz DJ, Ranlett G. Can communication deviance be measured in a family problem-solving interaction? Fam Process. 1990;29:213–226 [DOI] [PubMed] [Google Scholar]
- 36. Kymalainen JA, Weisman AG, Rosales GA, Armesto JC. Ethnicity, expressed emotion, and communication deviance in family members of patients with schizophrenia. J Nerv Ment Dis. 2006;194:391–396 [DOI] [PubMed] [Google Scholar]
- 37. Wynne LC, Cole RE, Perkins P. University of Rochester child and family study: risk research in progress. Schizophr Bull. 1987;13:463–476 [DOI] [PubMed] [Google Scholar]
- 38. Siira V, Wahlberg KE, Hakko H, Laksy K, Tienari P. Interaction of genetic vulnerability to schizophrenia and communication deviance of adoptive parents associated with MMPI schizophrenia vulnerability indicators of adoptees. Nord J Psychiatry. 2007;61:418–426 [DOI] [PubMed] [Google Scholar]
- 39. Metsanen M, Wahlberg KE, Saarento O, et al. Stability of Thought Disorder Index among high-risk and low-risk adoptees in the Finnish adoptive family study of schizophrenia. Eur Psychiatry. 2005;20:35–40 [DOI] [PubMed] [Google Scholar]
- 40. Roisko R, Wahlberg KE, Hakko H, Wynne L, Tienari P. Communication deviance in parents of families with adoptees at a high or low risk of schizophrenia-spectrum disorders and its associations with attributes of the adoptee and the adoptive parents. Psychiatry Res. 2011;185:66–71 [DOI] [PubMed] [Google Scholar]
- 41. Wahlberg KE, Wynne LC, Oja H, et al. Gene-environment interaction in vulnerability to schizophrenia: findings from the Finnish adoptive family study of schizophrenia. Am J Psychiatry. 1997;154:355–362 [DOI] [PubMed] [Google Scholar]
- 42. Wahlberg KE, Wynne LC, Oja H, et al. Thought disorder index of Finnish adoptees and communication deviance of their adoptive parents. Psychol Med. 2000;30:127–136 [DOI] [PubMed] [Google Scholar]
- 43. Wynne L, Singer MT, Bartko J, Toohey M. Schizophrenics and their families: research on parental communication. In: Tanner JM, ed. Developments in Psychiatric Research. London, UK: Hodder & Stoughton; 1977:254–286 [Google Scholar]
- 44. Wynne LC. The epigenesis of relational systems: a revised developmental perspective. Adolesc Psychiatry. 1993;19:240–254 [PubMed] [Google Scholar]
- 45. Velligan DI, Funderburg LG, Giesecke SL, Alexander AL. Longitudinal analysis of communication deviance in the families of schizophrenic patients. Psychiatry. 1995;58:6–19 [DOI] [PubMed] [Google Scholar]
- 46. Velligan DI, Miller AL, Eckert SL, et al. The relationship between parental communication deviance and relapse in schizophrenic patients in the 1-year period after hospital discharge: a pilot study. J Nerv Ment Dis. 1996;184:490–496 [DOI] [PubMed] [Google Scholar]
- 47. Lewis JM, Rodnick EH, Goldstein MJ. Intrafamilial interactive behavior, parental communication deviance, and risk for schizophrenia. J Abnorm Psychol. 1981;90:448–457 [DOI] [PubMed] [Google Scholar]
- 48. Lieber DJ. Parental focus of attention in a videotape feedback task as a function of hypothesized risk for offspring schizophrenia. Fam Process. 1977;16:467–475 [DOI] [PubMed] [Google Scholar]
- 49. Rund BR. Attention, communication, and schizophrenia. Yale J Biol Med. 1985;58:265–273 [PMC free article] [PubMed] [Google Scholar]
- 50. Wild CM, Shapiro LN, Goldenberg L. Transactional communication disturbances in families of male schizophrenics. Fam Process. 1975;14:131–160 [Google Scholar]
- 51. Docherty NM, Gordinier SW, Hall MJ, Cutting LP. Communication disturbances in relatives beyond the age of risk for schizophrenia and their associations with symptoms in patients. Schizophr Bull. 1999;25:851–862 [DOI] [PubMed] [Google Scholar]
- 52. Wild C, Singer M, Rosman B, Ricci J, Lidz T. Measuring disordered styles of thinking using object sorting test on parents of schizophrenic patients. Arch Gen Psychiatry. 1965;13:471–476 [DOI] [PubMed] [Google Scholar]
- 53. Palombo SR, Merrifield J, Weigert W, Morris GO, Wynne LC. Recognition of parents of schizophrenics from excerpts of family therapy interviews. Psychiatry. 1967;30:405–412 [DOI] [PubMed] [Google Scholar]
- 54. Morris GO, Wynne LC. Schizophrenic offspring and parental styles of communication : a predictive study using excerpts of family therapy recordings. Psychiatry. 1965;28:19–44 [DOI] [PubMed] [Google Scholar]
- 55. Docherty NM. Communication deviance, attention, and schizotypy in parents of schizophrenic patients. J Nerv Ment Dis. 1993;181:750–756 [DOI] [PubMed] [Google Scholar]
- 56. Docherty NM, Sledge WH, Wexler BE. Affective reactivity of language in stable schizophrenic outpatients and their parents. J Nerv Ment Dis. 1994;182:313–318 [DOI] [PubMed] [Google Scholar]
- 57. Docherty NM, Miller TN, Lewis MA. Communication disturbances in the natural speech of schizophrenic patients and non-schizophrenic parents of patients. Acta Psychiatr Scand. 1997;95:500–507 [DOI] [PubMed] [Google Scholar]
- 58. Sass LA, Gunderson JG, Singer MT, Wynne LC. Parental communication deviance and forms of thinking in male schizophrenic offspring. J Nerv Ment Dis. 1984;172:513–520 [DOI] [PubMed] [Google Scholar]
- 59. Wagener DK, Hogarty GE, Goldstein MJ, Asarnow RF, Browne A. Information processing and communication deviance in schizophrenic patients and their mothers. Psychiatry Res. 1986;18:365–377 [DOI] [PubMed] [Google Scholar]
- 60. Nuechterlein KH, Goldstein MJ, Ventura J, Dawson ME, Doane JA. Patient-environment relationships in schizophrenia: information processing, communication deviance, autonomic arousal, and stressful life events. Br J Psychiatry. 1989(suppl 5):84–89 [PubMed] [Google Scholar]
- 61. Rund BR, Oie M, Borchgrevink TS, Fjell A. Expressed emotion, communication deviance and schizophrenia. Psychopathology. 1995;28:220–228 [DOI] [PubMed] [Google Scholar]
- 62. Behrens MI, Rosenthal AJ, Chodoff P. Communication in lower class families of schizophrenics: II. Observations and findings. Arch Gen Psychiatry. 1968;18:689–696 [DOI] [PubMed] [Google Scholar]
- 63. Doane JA, Miklowitz DJ, Oranchak E, et al. Parental communication deviance and schizophrenia: a cross-cultural comparison of Mexican- and Anglo-Americans. J Abnorm Psychol. 1989;98:487–490 [DOI] [PubMed] [Google Scholar]
- 64. Carlson RG. Communication deviance, expressed emotion and family cohesion in schizophrenia [Unpublished doctoral dissertation]. Florida: University of Miami; 2011. [Google Scholar]
- 65. Docherty NM, Gordinier SW. Immediate memory, attention and communication disturbances in schizophrenia patients and their relatives. Psychol Med. 1999;29:189–197 [DOI] [PubMed] [Google Scholar]
- 66. Rund BR. Communication deviances in parents of schizophrenics. Fam Process. 1986;25:133–147 [DOI] [PubMed] [Google Scholar]
- 67. Solana RM. Communication disturbances in parents of schizophrenics. Acta Psychiatr Scand. 1988;77:427–434 [DOI] [PubMed] [Google Scholar]
- 68. Hirsch SR, Leff JP. Abnormalities in Parents of Schizophrenics: A Review of the Literature and an Investigation of Communication Defects and Deviances. New York, NY: Oxford University Press; 1975. [Google Scholar]
- 69. Liem JH. Family studies of schizophrenia: an update and commentary. Schizophr Bull. 1980;6:429–455 [DOI] [PubMed] [Google Scholar]
- 70. Jacob T. Family interaction in disturbed and normal families: a methodological and substantive review. Psychol Bull. 1975;82:33–65 [DOI] [PubMed] [Google Scholar]
- 71. Kymalainen JA, Weisman de Mamani AG. Expressed emotion, communication deviance, and culture in families of patients with schizophrenia: a review of the literature. Cultur Divers Ethnic Minor Psychol. 2008;14:85–91 [DOI] [PubMed] [Google Scholar]
- 72. Dell PF. The relationship between the communication of parents and sons in schizophrenic and normal families [Unpublished doctoral dissertation]. Austin: University of Texas; 1977. [Google Scholar]
- 73. Liem JH. Effects of verbal communications of parents and children: comparison of normal and schizophrenic families. J Consult Clin Psychol. 1974;42:438–450 [DOI] [PubMed] [Google Scholar]
- 74. Waxler NE. Parent and child effects on cognitive performance: experimental approach to etiological and responsive theories of schizophrenia. Fam Process. 1974;13:1–22 [Google Scholar]
- 75. Lidz T, Wild C, Schafer S, Rosman B, Fleck S. Thought disorders in the parents of schizophrenic patients: a study utilizing the object sorting test. J Psychiatr Res. 1962;1:193–200 [DOI] [PubMed] [Google Scholar]
- 76. Rosman B, Wild C, Ricci J, Fleck S, Lidz T. Thought disorders in the parents of schizophrenic patients: a further study utilizing the object sorting test. J Psychiatr Res. 1964;2:211–221 [DOI] [PubMed] [Google Scholar]
- 77. Catts SV, Mcconaghy N, Armstrong MS, Ward PB, Fox AM. High response threshold on the Lovibond-Rapaport object sorting test in the parents of schizophrenics. Pers Individ Dif. 1992;13:893–897 [Google Scholar]
- 78. Romney D. Psychometrically assessed thought disorder in schizophrenic and control patients and in their parents and siblings. Br J Psychiatry. 1969;115:999–1002 [DOI] [PubMed] [Google Scholar]
- 79. Schopler E, Loftin J. Thinking disorders in parents of young psychotic children. J Abnorm Psychol. 1969;74:281–287 [DOI] [PubMed] [Google Scholar]
- 80. Schopler E, Loftin J. Thought disorders in parents of psychotic children: a function of test anxiety. Arch Gen Psychiatry. 1969;20:174–181 [DOI] [PubMed] [Google Scholar]
- 81. Catts SV, McConaghy N, Ward PB, Fox AM, Hadzi-Pavlovic D. Allusive thinking in parents of schizophrenics : meta-analysis. J Nerv Ment Dis. 1993;181:298–302 [DOI] [PubMed] [Google Scholar]
- 82. Loveland NT, Wynne LC, Singer MT. The family Rorschach: a new method for studying family interaction. Fam Process. 1963;2:187–215 [Google Scholar]
- 83. Loveland N. The relation Rorschach: a technique for studying interaction. J Nerv Ment Dis. 1967;145:93–105 [DOI] [PubMed] [Google Scholar]
- 84. Singer MT. The consensus Rorschach and family transaction. J Proj Tech Pers Assess. 1968;32:348–351 [DOI] [PubMed] [Google Scholar]
- 85. Shapiro LN, Wild CM. Product of consensus Rorschach in families of male schizophrenics. Fam Process. 1976;15:211–224 [DOI] [PubMed] [Google Scholar]
- 86. Holte A, Wichstrøm L. Disconfirmatory feedback in families of schizophrenics. Scand J Psychol. 1990;31:198–211 [DOI] [PubMed] [Google Scholar]
- 87. Wichstrø, m L, Anderson AM, Holte A, Husby R, Wynne LC. Confirmatory and disconfirmatory family communication as predictor of offspring socio-emotional functioning : a 10- to 14-year follow-up of children at risk. Acta Psychiatr Scand. 1996;93:49–56 [DOI] [PubMed] [Google Scholar]
- 88. Wichstrøm L, Anderson AM, Holte A, Wynne LC. Disqualifying family communication and childhood social competence as predictors of offspring’s mental health and hospitalization : a 10- to 14-year longitudinal study of children at risk of psychopathology. J Nerv Ment Dis. 1996;184:581–588 [DOI] [PubMed] [Google Scholar]
- 89. Wichstrøom L, Holte A. Confirmatory and disconfirmatory feedback in families of schizophrenics, pathological controls and normals: replication and expansion. Acta Psychiatr Scand. 1991;84:575–578 [DOI] [PubMed] [Google Scholar]
- 90. Holte A, Wichstrøm L. Relationships between personality development, interpersonal perception and communication in parents of schizophrenics, psychiatric controls and normal subjects. Acta Psychiatr Scand. 1991;84:46–57 [DOI] [PubMed] [Google Scholar]
- 91. Wichstrøm L, Holte A, Wynne LC. Disqualifying family communication and anxiety in offspring at risk for psychopathology. Acta Psychiatr Scand. 1993;88:74–79 [DOI] [PubMed] [Google Scholar]
- 92. Wichstrøm L, Holte A, Husby R, Wynne LC. Competence in children at risk for psychopathology predicted from confirmatory and disconfirmatory family communication. Fam Process. 1993;32:203–220 [DOI] [PubMed] [Google Scholar]
- 93. Holte A, Wichstrø, m L. Confirmatory and disconfirmatory feedback in families of schizophrenics, pathological controls and normals. Acta Psychiatr Scand. 1990;81:477–482 [DOI] [PubMed] [Google Scholar]
- 94. Wichstrøm L, Holte A. Reciprocated self-disqualification among parents of schizophrenics. Acta Psychiatr Scand. 1992;86:201–206 [DOI] [PubMed] [Google Scholar]
- 95. Harrow M, Quinlan DM. Disordered Thinking and Schizophrenic Psychopathology. New York, NY: Gardner Press; 1985. [Google Scholar]
- 96. Haimo SF, Holzman PS. Thought disorder in schizophrenics and normal controls: social class and race differences. J Consult Clin Psychol. 1979;47:963–967 [DOI] [PubMed] [Google Scholar]
- 97. Docherty NM, Rhinewine JP, Labhart RP, Gordinier SW. Communication disturbances and family psychiatric history in parents of schizophrenic patients. J Nerv Ment Dis. 1998;186:761–768 [DOI] [PubMed] [Google Scholar]
- 98. Docherty NM, Hall MJ, Gordinier SW. Affective reactivity of speech in schizophrenia patients and their nonschizophrenic relatives. J Abnorm Psychol. 1998;107:461–467 [DOI] [PubMed] [Google Scholar]
- 99. Doane JA, West KL, Goldstein MJ, Rodnick EH, Jones JE. Parental communication deviance and affective style: predictors of subsequent schizophrenia spectrum disorders in vulnerable adolescents. Arch Gen Psychiatry. 1981;38:679–685 [DOI] [PubMed] [Google Scholar]
- 100. Albers LJ, Doane JA, Mintz J. Social competence and family environment: 15 year follow-up of disturbed adolescents. Fam Process. 1986;25:379–389 [DOI] [PubMed] [Google Scholar]
- 101. Asarnow JR, Goldstein MJ, Ben-Meir S. Parental communication deviance in childhood onset schizophrenia spectrum and depressive disorders. J Child Psychol Psychiatry. 1988;29:825–838 [DOI] [PubMed] [Google Scholar]
- 102. Glaser RT. Family interaction patterns and schizophrenia: a Rorschach study [Unpublished doctoral dissertation]. Berkeley: University of California, Berkeley; 1976. [Google Scholar]
- 103. Hirsch SR, Leff JP. Parental abnormalities of verbal communication in the transmission of schizophrenia. Psychol Med. 1971;1:118–127 [DOI] [PubMed] [Google Scholar]
- 104. Wender PH, Rosenthal D, Rainer JD, Greenhill L, Sarlin MB. Schizophrenics’ adopting parents: psychiatric status. Arch Gen Psychiatry. 1977;34:777–784 [DOI] [PubMed] [Google Scholar]
- 105. Wynne L. Family transactions and schizophrenia: II. Conceptual considerations for a research strate gy. In: Romano J, ed. The Origins of Schizophrenia. Amsterdam, The Netherlands: Excerpta Medica; 1967:165–178 [Google Scholar]
- 106. Biostat. Comprehensive Meta-analysis [computer program]. Version 2.2.064. Englewood, NJ: Biostat; 2005. [Google Scholar]
- 107. Hedges L. Distributional theory for Glass’ estimator of effect size and related estimators. J Educ Behav Stat. 1981;6:107–128 [Google Scholar]
- 108. Chinn S. A simple method for converting an odds ratio to effect size for use in meta-analysis. Stat Med. 2000;19:3127–3131 [DOI] [PubMed] [Google Scholar]
- 109. Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0. The Cochrane Collaboration; 2011. http://www.cochrane-handbook.org Accessed January 15, 2013. [Google Scholar]
- 110. Hedges LV. A random effects model for effect sizes. Psychol Bull. 1983;93:388–395 [Google Scholar]
- 111. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Erlbaum; 1988. [Google Scholar]
- 112. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994;50:1088–1101 [PubMed] [Google Scholar]
- 113. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315:629–634 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 114. Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics. 2000;56:455–463 [DOI] [PubMed] [Google Scholar]
- 115. Liem JH. Intrafamily communication and schizophrenic thought disorder: an etiologic or responsive relationship? Clin Psychol. 1976;29:28–30 [Google Scholar]
- 116. Covelman K. Speaker and listener deficits in mother-son dyads from schizophrenia and normal families [Unpublished doctoral dissertation]. Austin: University of Texas; 1977. [Google Scholar]
- 117. Nugter MA, Dingemans PMAJ, Linszen DH, Vander Does AJW, Gersons BPR. Parental communication deviance: its stability and the effect of family treatment in recent onset schizophrenia. Acta Psychiatr Scand. 1997;95:199–204 [DOI] [PubMed] [Google Scholar]
- 118. Wahlberg KE, Wynne LC, Keskitalo P, et al. Long-term stability of communication deviance. J Abnorm Psychol. 2001;110:443–448 [DOI] [PubMed] [Google Scholar]
- 119. Wearden AJ, Tarrier N, Barrowclough C, Zastowny TR, Rahill AA. A review of expressed emotion research in health care. Clin Psychol Rev. 2000;20:633–666 [DOI] [PubMed] [Google Scholar]
- 120. Miklowitz DJ. Family risk indicators in schizophrenia. Schizophr Bull. 1994;20:137–149 [DOI] [PubMed] [Google Scholar]
- 121. Hahlweg K, Goldstein MJ, Nuechterlein KH, et al. Expressed emotion and patient-relative interaction in families of recent onset schizophrenics. J Consult Clin Psychol. 1989;57:11–18 [DOI] [PubMed] [Google Scholar]
- 122. Kinney DK, Holzman PS, Jacobsen B, et al. Thought disorder in schizophrenic and control adoptees and their relatives. Arch Gen Psychiatry. 1997;54:475–479 [DOI] [PubMed] [Google Scholar]
- 123. Shenton ME, Solovay MR, Holzman PS, Coleman M, Gale HJ. Thought disorder in the relatives of psychotic patients. Arch Gen Psychiatry. 1989;46:897–901 [DOI] [PubMed] [Google Scholar]
- 124. Snitz BE, Macdonald AW, Carter CS. Cognitive deficits in unaffected first-degree relatives of schizophrenia patients: a meta-analytic review of putative endophenotypes. Schizophr Bull. 2006;32:179–194 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 125. Levy DL, Coleman MJ, Sung H, et al. The genetic basis of thought disorder and language and communication disturbances in schizophrenia. J Neurolinguistics. 2010;23:176. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 126. Tolosa A, Sanjuan J, Dagnall AM, Moltó MD, Herrero N, de Frutos R. FOXP2 gene and language impairment in schizophrenia: association and epigenetic studies. BMC Med Genet. July 22, 2010; 10.1186/1471-2350-11-114 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 127. Wang KS, Zhang QY, Liu XF, Wu LY, Zeng M. PKNOX2 is associated with formal thought disorder in schizophrenia: a meta-analysis of two genome-wide association studies. J Mol Neurosci. 2012;48:265–272 [DOI] [PubMed] [Google Scholar]
- 128. Hamilton EB, Jones M, Hammen C. Maternal interaction style in affective disordered, physically ill, and normal women. Fam Process. 1993;32:329–340 [DOI] [PubMed] [Google Scholar]
- 129. Wynne L, Singer MT, Toohey M. Communication of the adoptive parents of schizophrenics. In: Jørstad J, Ugelstad E, eds. Schizophrenia 75: Psychotherapy, Family Studies, Research. Oslo, Norway: Universitetsforlaget; 1976:413–452 [Google Scholar]
- 130. Reininghaus U, Priebe S, Bentall RP. Testing the psychopathology of psychosis: evidence for a general psychosis dimension. Schizophrenia Bull. January 18, 2012; 10.1093/schbul/sbr182 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 131. Demjaha A, Morgan K, Morgan C, et al. Combining dimensional and categorical representation of psychosis: the way forward for DSM-V and ICD-11? Psychol Med. 2009;39:1943–1955 [DOI] [PubMed] [Google Scholar]
- 132. Velligan D, Christensen A, Goldstein MJ, Margolin G. Parental communication deviance: its relationship to parent, child, and family system variables. Psychiatry Res. 1988;26:313–325 [DOI] [PubMed] [Google Scholar]
- 133. Vygotsky L. Thought and Language. Cambridge, MA: MIT Press; 1962. [Google Scholar]
- 134. Wynne L. Methodologic and conceptual issues in the study of schizophrenics and their families. J Psychiatr Res. 1968;6:185–1995731854 [Google Scholar]
- 135. Goldsmith DF, Rogoff B. Mothers’ and toddlers’ coordinated joint focus of attention: variations with maternal dysphoric symptoms. Dev Psychol. 1997;33:113–119 [DOI] [PubMed] [Google Scholar]
- 136. Hustedt JT, Raver CC. Scaffolding in low-income mother-child dyads: relations with joint attention and dyadic reciprocity. Int J Behav Dev. 2002;26:113–119 [Google Scholar]
- 137. Miller JL, Ables EM, King AP, West MJ. Different patterns of contingent stimulation differentially affect attention span in prelinguistic infants. Infant Behav Dev. 2009;32:254–261 [DOI] [PubMed] [Google Scholar]
- 138. Hobson RP, Patrick MPH, Crandell LE, García Pérez RM, Lee A. Maternal sensitivity and infant triadic communication. J Child Psychol Psychiatry. 2004;45:470–480 [DOI] [PubMed] [Google Scholar]
- 139. Carpenter M, Nagell K, Tomasello M. Social cognition, joint attention, and communicative competence from 9 to 15 months of age. Monogr Soc Res Child Dev. 1998;63:1–143 [PubMed] [Google Scholar]
- 140. Mundy P, Jarrold W. Infant joint attention, neural networks and social cognition. Neural Netw. 2010;23:985–997 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 141. Scaife M, Bruner JS. Capacity for joint visual attention in the infant. Nature. 1975;253:265–266 [DOI] [PubMed] [Google Scholar]
- 142. Tomasello M, Farrar MJ. Joint attention and early language. Child Dev. 1986;57:1454–1463 [PubMed] [Google Scholar]
- 143. Tasker SL, Schmidt LA. The “dual usage problem” in the explanations of “joint attention” and children’s socioemotional development: a reconceptualization. Dev Rev. 2008;28:263–288 [Google Scholar]
- 144. Bearden CE, Rosso IM, Hollister JM, Sanchez LE, Hadley T, Cannon TD. A prospective cohort study of childhood behavioral deviance and language abnormalities as predictors of adult schizophrenia. Schizophr Bull. 2000;26:395–410 [DOI] [PubMed] [Google Scholar]
- 145. Cannon M, Caspi A, Moffitt TE, et al. Evidence for early-childhood, pan-developmental impairment specific to schizophreniform disorder: results from a longitudinal birth cohort. Arch Gen Psychiatry. 2002;59:449–456 [DOI] [PubMed] [Google Scholar]
- 146. Jones P, Rodgers B, Murray R, Marmot M. Child developmental risk-factors for adult schizophrenia in the British 1946 birth cohort. Lancet. 1994;344:1398–1402 [DOI] [PubMed] [Google Scholar]
- 147. Bird V, Premkumar P, Kendall T, Whittington C, Mitchell J, Kuipers E. Early intervention services, cognitive-behavioural therapy and family intervention in early psychosis: systematic review. Br J Psychiatry. 2010;197:350–356 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 148. Pharoah F, Mari J, Rathbone J, Wong W. Family Intervention for Schizophrenia (Review). Cochrane Database of Systematic Reviews. Oxford, UK: John Wiley & Sons; 2006. [DOI] [PubMed] [Google Scholar]
- 149. Pilling S, Bebbington P, Kuipers E, et al. Psychological treatments in schizophrenia: I. Meta-analysis of family intervention and cognitive behaviour therapy. Psychol Med. 2002;32:763–782 [DOI] [PubMed] [Google Scholar]