Schizophrenia Research 140 (2012) 41–45

Contents lists available at SciVerse ScienceDirect

Schizophrenia Research

journal homepage: www.elsevier.com/locate/schres

On the boundaries of blunt affect/alogia across severe mental illness: Implications for Research Domain Criteria

Alex S. Cohen a,⁎, Gina M. Najolia a, Yunjung Kim b, Thomas J. Dinzeo c a Louisiana State University, Department of Psychology, USA b Louisiana State University, Department of Communication Sciences and Disorders, USA c Rowan University, Department of Psychology, USA article info abstract

Article history: There is growing awareness that reduced expressive behaviors (e.g., blunt affect, alogia, psychomotor retar- Received 16 April 2012 dation) are characteristic of a range of psychiatric conditions, including mood and schizophrenia-spectrum Received in revised form 27 June 2012 disorders. From a Research Domain Criteria (RDoC) perspective, it would be critical to determine whether Accepted 2 July 2012 these symptoms manifest similarly across diagnostic groups — as they may share common pathophysiolog- Available online 23 July 2012 ical underpinnings. The present study employed computerized acoustic analysis of speech produced in reac- tion to a range of visual stimuli in 48 stable outpatients with schizophrenia and mood disorders to offer Keywords: Schizophrenia preliminary understanding of this issue. Speaking assessments were administered 1 week-apart to examine Affect how temporal stability might vary as a function of clinical diagnosis and symptom severity. Speech character- Negative istics generally did not differ between groups and were similarly, and for the most part, highly stable over Deficit time. Aspects of speech were significantly associated with severity of psychosis and negative symptoms, Cognitive but not with clinical depression/anxiety severity. Moreover, stability of speech characteristics generally did Mood not vary as a function of diagnostic group or clinical severity. The magnitudes of group differences were al- most exclusively in the negligible to small range. Speech production was associated with social functioning deficits. In sum, these preliminary data suggest that speech variables tap a stable and clinically important facet of psychopathology that cut across diagnostic categories. Computerized acoustic analysis of speech ap- pears to be a promising method for understanding the pathological manifestation of these variables. © 2012 Published by Elsevier B.V.

1. Introduction constricted affect in subclinical form, these expressive deficits are diag- nostic criteria of both major depression and schizophrenia-spectrum There is growing awareness that psychopathological processes tran- disorders (American Psychiatric Association (APA), 1994). Moreover, scend psychiatric diagnoses. Of note, genetic, molecular, anatomical, be- expressive deficits are a component of schizotypy — defined as the havioral and subjective pathological phenomena are rarely constrained personality organization reflecting a putatively genetic vulnerability to to a single psychiatric disorder. The Research Domain Criteria (RDoC), schizophrenia-spectrum disorders thought to occur in approximately advanced by the National Institute of Mental Health (NIMH), is a 10% of the population (Meehl, 1962; Collins et al., 2005; Cohen and novel approach to understanding psychopathology that focuses on the Hong, 2011). To date, expressive deficits have been primarily studied discovery of identifiable subtypes across rather than within mental dis- in the context of schizophrenia negative symptoms. However, growing orders (Insel et al., 2010). Initial efforts have focused on five broad clin- empirical attention has been paid to their presence in major depression ical phenotype categories (e.g., negative valence systems, cognitive as well. Interestingly, several studies have provided evidence that se- systems). The present study sought to complement this effort by verity of clinically-rated blunted affect is similar in schizophrenia and conducting a preliminary evaluation of a novel domain, involving a re- major depression (Kulhara and Chadda, 1987; Tremeau et al., 2005). duction in expressive behavior, as a potential Research Domain Criteria Expressive deficits are also relatively stable and associated with a host for severe mental illness. of deleterious variables, including poor functioning and prognosis in A reduction in expressive behaviors appears to be a prominent individuals with schizophrenia-spectrum disorders (Fenton and feature of a range of severe psychiatric conditions. Observable as McGlashan, 1991; Mueser et al., 1994; Gur et al., 2006) and major blunt/flat affect, alogia, or psychomotor retardation in clinical form, or depression (Parker et al., 1992). Moreover, clinically-rated blunted affect has been similarly related to social skill deficits for both disorders (Herbener and Harrow, 2004; Mueser et al., 2010), suggesting that ⁎ Corresponding author at: Louisiana State University, Department of Psychology, fi 236 Audubon Hall, Baton Rouge, LA 70803, USA. Tel.: +1 225 578 7017. expressive de cits have common pathological sequelae in these two E-mail address: [email protected] (A.S. Cohen). disorders.

0920-9964/$ – see front matter © 2012 Published by Elsevier B.V. doi:10.1016/j.schres.2012.07.001 42 A.S. Cohen et al. / Schizophrenia Research 140 (2012) 41–45

As yet, our understanding of expressive characteristics is limited, Table 1 in large part, due to the reliance on Likert-type clinical rating scales Descriptive statistics for demographic and clinical variables for the mood disorder and schizophrenia groups. (e.g., SANS; Andreasen, 1984) for measurement. These instruments are not ideal for understanding expression due to their restricted Mood disorder group Schizophrenia group scoring range and their use of vaguely defined anchors within an or- (n=22) (n=26) dinal response format (Lader, 2000; Alpert et al., 2002; Cohen et al., % Caucasian 64% (n=14) 35% (n=9) 2008). For the past 7 years, our laboratory has been employing com- African-American 36% (n=8) 65% (n=17) % Male 60% (n=13) 62% (n=16) puterized acoustic analysis of natural speech to improve the under- a fi Current medications standing of expressive de cits. The use of a sensitive analytic 2nd generation antipsychotic 71% (n=12) 76% (n=19) approach that is behaviorally-based offers improved sensitivity and 1st generation antipsychotic 12% (n=2) 16% (n=4) reliability over clinical rating scales (Lader, 2000; Alpert et al., Antidepressants 47% (n=8) 40% (n=10) 2002; Cohen et al., 2008). The primary aim of this project was to ex- Mood stabilizers 12% (n=2) 32% (n=8) Anticholinergics 6% (n=1) 20% (n=5) amine patients with psychotic, depressed and manic symptoms in Psychiatric history acoustic-based measures of expressivity. We analyzed natural Major depression 100% (n=22) 46% (n=12) speech procured from a wide range of affectively-valenced speaking Manic episodes 41% (n=9) 27% (n=7) tasks, administered across two testing sessions scheduled a week Psychosis symptoms 27% (n=6) 100% (n=26) apart. We used our computer-based technology in patients to an- Current diagnoses Major depressive episode 20% (n=5) 20% (n=4) swer the following questions: 1) To what degree are speech charac- Manic episodes 9% (n=2) 8% (n=2) teristics stable over a week epoch? 2) To what degree do speech characteristics reflect clinical state (i.e., anxiety/depression, psycho- Mean SD Mean SD sis or mania) or diagnostic history (i.e., history of Bipolar I, depres- Father's educational level 10.00 5.24 11.79 3.22 sion, psychosis), and how does this stability differ between Number hospitalizations 4.23 4.69 4.55 4.80 patients with schizophrenia and those with mood disorders? 3) GAF 51.05 8.97 47.50 9.44 SOFAS 52.33 8.70 50.11 10.02 Across patients, to what degree are speech characteristics associated Reading ability (WRAT) 55.72 9.04 53.16 7.94 with social functioning deficits? Age 46.17 9.69 39.87 9.96 Brief Psychiatric Rating Scale factor scores: 2. Methods Mania/excitement 10.26 5.47 10.39 4.63 Negative 6.78 2.98 8.51 3.70 Positive 7.78 3.10 12.26 5.30 2.1. Subjects Depression/anxiety 12.27 5.18 10.31 5.13

a Note — missing data for 5 mood disorder and 1 schizophrenia patients. Subjects were outpatients at a community mental health hospital (n=48). These subjects included 26 patients with Diagnostic and Statistical Manual of Mental Disorders 4th edition (DSM-IV; APA, 1994) diagnosed schizophrenia and 22 patients with a history of a DSM-IV major depressive episode without a history of schizophrenia- suspiciousness, unusual thought content, disorientation, and hallucina- spectrum disorder. Note that 10 of the individuals in the mood disorder tion items), negative (i.e., self-neglect, blunted affect, motor retardation, – group also met criteria for a manic episode at some point in their lives. and emotional withdrawal items), depression anxiety (i.e., depression, Thus, 45% of the mood disorder sample met criteria for bipolar disorder, guilt, suicidality, and anxiety items), and mania/excitement (i.e., motor whereas 55% met criteria for unipolar major depressive disorder. Demo- hyperactivity, elevated mood, excitement, distractibility, hostility, and fi graphic and clinical information is provided in Table 1.Diagnoseswere grandiosity items) symptoms (de ned in Ventura et al., 2000)were made based on information obtained from the patients' medical records employed here. Preliminary diagnoses and ratings were made by one of and from a structured clinical interview (SCID; First et al., 1996). Exclu- four doctoral level students who were trained to criterion (Intra-class fi sion criteria included: a) a Global Assessment of Functioning rating Correlation Coef cient (ICC) values>.70). All diagnoses and ratings (GAF; APA, 1994) below 30, indicating a level of impairment that were videotaped and reviewed during a case conference meeting that could interfere with participation in the study, b) documented evidence was led by a licensed clinical psychologist with considerable diagnostic of intellectual disability from the medical records, c) current or histori- experience (Alex S. Cohen), and were recorded only when full agreement cal DSM-IV diagnosis of alcohol or drug abuse suggestive of severe bythecaseconferencememberswasmade. physiological symptoms (i.e., delirium tremens, repeated loss of con- sciousness), and d) history of significant head trauma (requiring over- 2.3. Verbal achievement night hospitalization). All patients were clinically stable (GAF score>30) at the time of testing and were receiving pharmacotherapy To examine potential confounds associated with individual differ- under the supervision of a multi-disciplinary team. Patients received ences in verbal achievement, we administered the word reading $40 for participation in this study. This study was approved by the ap- subtest of the Wide Range Achievement Test 4th edition (WRAT-4; propriate Human Subject Review Boards and all participants offered in- Wilkinson and Robertson, 2006) to participants. formed consent prior to participating in the study. For additional information about recruitment, subjects and methods, the reader is re- 2.4. Social functioning ferred to Cohen et al. (under review). Social functioning was measured using the Social and Occupation- 2.2. Diagnostic and symptom ratings al Functioning and Assessment Scale (SOFAS) from the DSM-IV (APA, 1994), a scale from 1 to 100 reflecting social and occupational func- The Brief Psychiatric Rating Scale (BPRS; Lukoff et al., 1986)wasse- tioning during the prior month. lected to characterize patients' symptoms for this study because it is a broad measure of psychopathology. BPRS ratings were based on informa- 2.5. Speech tasks tion obtained from medical records, the patients' treatment teams, and self-report and behavioral observations made during the research inter- Participants were asked to verbalize their reactions to pictures view. Factor subscale scores reflecting positive (i.e., bizarre behavior, from the International Affective Picture System (IAPS; Lang et al., A.S. Cohen et al. / Schizophrenia Research 140 (2012) 41–45 43

2005) in six separate conditions grouped by emotional valence of major depressive disorder, mania and psychosis disorders to patients (i.e., good, bad, neutral) and arousal (i.e., high, low). Each condi- without a history of these disorders (regardless of comorbid psychopa- tion featured three separate IAPS images displayed for 20 s each. thology) on composite speech characteristics, and b) computing Administrations, which employed different images, were adminis- correlations between BPRS mania, negative, depression/anxiety and tered twice — separated approximately by a week epoch. The order positive symptom factors and SOFAS scores with composite speech of these administrations was counterbalanced across subjects. Pictures characteristics within the total group. Nonparametric statistics were were selected for their relative representation of their respective va- used for analyses involving change scores to address potential abnormal lence and arousal based on existing norms (Lang et al., 2005). Picture distributions. Cohen's d, interpreted as negligible (db.20), small (d= and recording onset and offset were controlled by experimental soft- .20–.49), medium (d=.50–.79) and large (d>.80), was reported ware. In this regard, each speaking condition was exactly 60 s. Block when appropriate to aid in interpretation, particularly in the case of order and picture order within each block were random. Blocks were null findings. All analyses in this study are two-tailed and all variables separated by a 30 s interval during which participants were instructed are normally distributed unless otherwise stated. Two extreme values to “relax and breathe deeply”.Inall,360sofspeechwereavailable (i.e., z-score>3.5 above the mean) for the Pause x variable were for each time point. Subjects were asked to discuss how the picture re- replaced with a score at 3.5 standard deviations above the mean. lates to them, what it means to them, what it reminds them of and how it makes them feel. Subjects were encouraged to speak for the full re- 3. Results cording time, and interviewers were not allowed to speak during this time. 3.1. Demographic and descriptive variables

2.6. Acoustic analysis Means and standard deviations of the descriptive and clinical variables were separately computed for the schizophrenia and The Computerized assessment of Affect from Natural Speech (CANS) mood disorder groups. These data are presented in Table 1. Group- protocol (Cohen et al., 2009, 2010), developed by our lab to assess vocal wise comparisons, using a series of ANOVA and Chi-square analyses expression, was employed. Speech was digitally recorded at 16 bits per suggested that the schizophrenia and mood disorder groups were second at a sampling frequency of 44,100 Hz using headset microphones. not statistically dissimilar for most of the variables, including psychi- The digitized recordings were analyzed using PRAAT (Boersma and atric history (i.e., history of manic episodes, functioning and reading Weenink, 2006), a program that has been used extensively in speech pa- ability; p's>.05). The groups were also similar in terms of manic- thology and linguistic studies. The PRAAT system organizes sound files excitement, negative and depression/anxiety subscale scores from into “frames” for analysis, which for the present study was set at a rate the BPRS. The schizophrenia group had a significantly lower history of of 100 per second. During each of these frames, frequency and volume major depressive episode (x2 [1]=.14.84, pb.001), a greater severity are quantified. Four measures were computed for this study, including: of psychosis symptoms (x2 [1]=30.27, pb.001), a higher percentage 2 average pause length (Pause x) — computed as the average time in of African-Americans (x [1]=6.20, pb.05) and a lower age (t [46]= seconds between utterances (defined as speech bounded by silence in ex- 2.72, pb.01). To address potential confounding effects of ethnicity and cess of 50 ms), number of utterances (UtteranceN) — computed as the age, we examined the relationships between these variables and the number of utterances within a speech sample, inflection (F0sd of sd) — speech characteristics. Speech characteristics were not significantly cor- computed as the standard deviation of the standard deviation values of related to age and were not significantly different between ethnic fundamental frequency determined within each individual utterance, groups (p's>.05). All analyses presented in this study were rerun con- and emphasis (dBsd of sd) — computed as the standard deviation of stan- trolling for the effects of age and ethnicity. Unless noted, there were dard deviation values of intensity determined within each individual ut- no substantive changes. terance. The average pause length and number of utterance variables conceptually tap into alogia (i.e., language production) whereas the 3.2. Group comparisons on expressivity inflection and emphasis variables tap into blunted affect. Table 2 contains the descriptive data for speech variables. With 2.7. Analyses only a few notable exceptions, the condition effects, time effects and group-based interactions were not statistically significant (p's>.10). The analyses were conducted in five steps. First, we computed and The schizophrenia versus mood disorder group had significantly compared descriptive and clinical variables between the patient higher F0sd of sd values (F[1, 46]=4.33, p=.04). Post-hoc t-tests groups (i.e., schizophrenia versus mood disorders) to identify poten- suggested that the schizophrenia group was significantly higher in tial confounding variables. Second, we compared the groups in F0sd of sd values for the second (t [46]=2.08, p=.04, d=.66) but speech characteristics across the arousal (i.e., high, low) and valence not for the first (t [46]=1.85, p=.07, d=.57) administrations com- (good, bad, neutral) conditions across the two time points using pared to the mood disorder group, though these effect size values mixed-model ANOVAs. We were interested in the group effects and were very similar. When either ethnicity or age was entered as a co- the group interactions, as these effects regard how speech characteristics variate, these differences became non-significant (p's>.10). The differ between diagnostic groups. Third, we examined the temporal sta- arousal by group interaction for UtteranceN was statistically signifi- bility of the CANS variables across the two speech samples administered cant (4.13[46]=4.13, p=.048). Post-hoc analysis revealed that pa- across the week epoch using Intra-class Correlation Coefficient values, tients with schizophrenia expressed fewer utterances in the low interpreted as excellent (ICC>.75), fair to good (ICC .40–.75) and poor versus high (mean and standard deviation [M±SD]=22.98±7.94 (ICCb.40) (Fleiss, 1986). Fourth, we computed zero-order correlations and 24.24±7.56 respectively) arousal conditions compared to pa- between the CANS variables (averaged across the two time points) to ex- tients with mood disorders (M±SD=24.06±9.25 and 24.05± plore how they are inter-related. Finally, we examined how individual 9.95) respectively (pb.05). This value also became non-significant differences in depression–anxiety, mania, negative and psychosis sever- when either ethnicity or age was entered as a covariate (p>.10). ity and social functioning were related to speech characteristics (aver- There were no significant effects of valence on speech characteristics. aged across the valence and arousal conditions across the two The lack of group effects (with the possible exception of F0sd of sd) assessments) and to change in speech characteristics (computed as suggests that speech characteristics were similar across groups. The the absolute values of change scores from T2−T1, averaged across va- lack of time effects suggests that speech characteristics were not lence and arousal conditions) by: a) comparing patients with a history demonstrably different across the two assessments. The small to 44 A.S. Cohen et al. / Schizophrenia Research 140 (2012) 41–45

Table 2 Table 3 Means and standard deviations for speech characteristics for the mood disorder and Zero-order correlations between speech characteristics (averaged across T1 and T2). schizophrenia groups across the various valence and arousal conditions. Pausex UtteranceN F0sd of sd dBsd of sd a b Speech condition Pausex UtteranceN F0sd of sd dBsd of sd Pausex 1 ––– ⁎⁎⁎ Schizophrenia group UtteranceN −.80 1 ––

Bad–high Baseline 4.08 (4.19) 24.46 (9.13) .22 (.11) 1.43 (.29) F0sd of sd .18 −.07 1 –

Follow-up 3.52 (1.41) 23.26 (7.24) .19 (.11) 1.39 (.33) dBsd of sd −.18 .25 .03 1 Bad–low Baseline 5.13 (7.27) 22.94 (8.89) .20 (.11) 1.40 (.27) ⁎⁎⁎pb.001. Follow-up 5.01 (8.24) 23.33 (7.25) .21 (.13) 1.31 (.34) Good–high Baseline 3.73 (3.12) 25.73 (7.90) .20 (.09) 1.48 (.32) Follow-up 3.25 (.91) 24.18 (5.89) .23 (.12) 1.32 (.27) Bivariatecorrelations,computedbetweentheBPRSfactorsand Good–low Baseline 3.89 (2.60) 23.05 (8.43) .23 (.12) 1.37 (.26) speech characteristics were computed (see Table 4). There were Follow-up 3.63 (1.28) 22.18 (5.58) .22 (.13) 1.33 (.29) five notable findings. First, both negative and psychosis symptom Neutral–high Baseline 4.15 (3.37) 23.38 (8.16) .20 (.12) 1.45 (.31) severity scores were significantly associated with increased paused Follow-up 3.40 (2.06) 24.39 (7.00) .22 (.13) 1.38 (.38) Neutral–low Baseline 3.79 (2.47) 25.05 (9.41) .22 (.10) 1.41 (.27) and decreased number of utterances. Interestingly, psychosis Follow-up 5.22 (6.23) 21.33 (8.03) .20 (.12) 1.35 (.31) severity was also associated with increased inflection. Second, mania–excitement scores were significantly associated with in- Mood disorder group creased inflection at a medium effect size level, but none of the Bad–high Baseline 4.11 (3.70) 24.05 (8.35) .15 (.06) 1.38 (.35) Follow-up 4.08 (2.88) 23.83 (9.81) .15 (.06) 1.27 (.27) other variables (small to negligible effect size levels). Third, de- Bad–low Baseline 4.46 (4.21) 24.36 (9.76) .15 (.06) 1.28 (.34) pression–anxiety severity was not significantly associated with Follow-up 3.86 (2.23) 24.49 (9.26) .14 (.08) 1.21 (.24) any of the speech characteristics (small to negligible effects). Final- Good–high Baseline 3.75 (2.42) 25.00 (9.05) .15 (.07) 1.41 (.33) ly, poorer social functioning was significantly associated with Follow-up 4.51 (2.64) 22.45 (10.82) .14 (.06) 1.28 (.25) fewer utterances and, at a trend level, longer pauses. There were Good–low Baseline 3.88 (2.77) 24.88 (9.01) .17 (.08) 1.38 (.29) Follow-up 3.63 (1.80) 24.91 (9.70) .16 (.07) 1.28 (.30) no substantive changes to these findings when we recomputed Neutral–high Baseline 4.69 (5.53) 24.65 (10.34) .17 (.07) 1.32 (.37) these correlations controlling for diagnosis. Follow-up 4.17 (3.21) 24.33 (11.30) .17 (.08) 1.28 (.24) With respect to potential clinical correlates of change in speech – Neutral low Baseline 4.08 (3.21) 24.11 (9.00) .18 (.09) 1.39 (.26) characteristics (computed as a difference score of the two assess- Follow-up 4.42 (2.49) 21.60 (8.74) .18 (.08) 1.30 (.28) ments), there were no differences in speech characteristics between a In seconds. patients with a history of major depression, manic episodes or psy- b Multiplied by a factor of 100 for presentation purposes. chosis relative to their respective comparison groups. Correlations between the BPRS factors and change in speech characteristics (see negligible effect size values suggest that lack of statistical significance Table 4) revealed few significant findings. Increased variability in was not a power issue. pause time across the two assessments was associated with more se- vere negative symptoms (at a trend level), more significant psychosis 3.3. Temporal stability (significant) and poorer social functioning (at a trend level). There were no substantive changes to these findings when we recomputed ICC values between the speech measures across the two time these correlations controlling for diagnosis. points were computed. Values for the mood disorder and schizophre- nia groups were as follows: Pausex (r's=.86 and .73 respectively), UtteranceN (r's=.93 and .82 respectively), F0sd of sd (r's=.89 and 4. Discussion .96 respectively) and dBsd of sd (r's=.63 and .40 respectively). The stability of speech characteristics ranged from fair to excellent, with The present study employed computerized analysis of natural higher stability seen in variables pertaining to speech production speech to understand expressivity in patients with a range of se- (i.e., alogia) and inflection and lower stability seen in variables vere mental disorders. There were four critical findings from this pertaining to emphasis. study. First, patients with schizophrenia and those with mood disorders did not, by and large, significantly differ in speech 3.4. Zero-order correlations characteristics across a range of affective and arousal conditions.

Results (Table 3) suggest that Pause x and UtteranceN variables were highly correlated with each other. The other variables were rel- Table 4 atively independent of each other. Bivariate correlations between clinical characteristics and speech characteristics.

Pausex UtteranceN F0sd of sd dBsd of sd 3.5. The relationship between psychiatric symptoms and expressivity Speech characteristics (averaged over T1 and T2) ⁎ Mania–excitement .02 .12 .23 .38 ⁎⁎ ⁎⁎ With respect to potential clinical correlates of speech characteristics Negative .47 −.55 .08 −.19 ⁎⁎ ⁎ ⁎ (aggregated across the two assessments), group comparisons revealed Psychosis .39 −.32 .33 .12 that, with only one exception (see below), there were no statistically Depression–anxiety −.07 .03 .17 −.19 + ⁎ Social functioning −.29 .36 −.12 −.04 significant differences in acoustic-based speech characteristics between a) patients with a history of major depression (n=34) and those with- Δ speech characteristics (T1−T2)a out (n=14) (p's>.35, d'sb.31), b) patients with a history of Bipolar I Mania–excitement −.10 −.18 .04 .06 disorder (n=16) and those without (n=32) (p's>.34, d'sb.29), or pa- Negative .28+ .17 −.21 −.12 ⁎ − tients with a history of psychosis (n=32) and those without (n=16) Psychosis .32 .20 .05 .22 Depression–anxiety −.08 −.08 −.05 .24 (raw data omitted for space concerns). Patients with a history of psy- Social functioning −.27+ −.07 .10 −.11 chosis had a trend for more inflection than those without (t [46]= ⁎ pb.05. 1.86, p=.07, d=.66; all other p'sb16, d'sb.44). With this one excep- ⁎⁎ pb.01. tion, the small to negligible effect size values suggest that lack of statis- +pb.10. tical significance was not a power issue. a Spearman's correlations. A.S. Cohen et al. / Schizophrenia Research 140 (2012) 41–45 45

Schizophrenia patients did evidence more inflection than those Contributors with mood disorders, though this difference was non-significant Alex S. Cohen was the primary investigator for this project and designed the study and wrote the bulk of the manuscript. Yunjung Kim, Gina M. Najolia and Thomas when demographic factors were controlled for. The magnitudes Dinzeo helped manage the literature searches and the analyses and provided concep- of these group differences were in the small to negligible range, tual material to the planning and presentation of this project. All authors contributed suggesting this was not a power issue. Second, speech characteris- to and have approved the final manuscript. tics were stable over a week epoch, and they did not, for the most part, vary as a function of clinical characteristics. Third, negative Conflict of interest There are no conflicts of interest to report. and psychosis symptom severity was significantly associated with reduced speech production and was, in the case of psychosis Acknowledgments fl symptoms, related to higher levels of in ection.Finally,aspects The authors wish to acknowledge the efforts of S Lee Hong, Neila Donovan, Melissa of speech production were associated with social dysfunctions, Beck, Jason Hicks and Sean Lane for their advice and guidance on this project. We providing further evidence that expressive characteristics are by would also like to thank the subjects for their participation. no means benign (e.g., Fenton and McGlashan, 1991). The chief finding from this study is that speech characteristics did References not change as a function of schizophrenia, major depression or Alpert, M., Shaw, R.J., Pouget, E.R., Lim, K.O., 2002. A comparison of clinical ratings with Bipolar I diagnosis or symptoms. These findings support the notion vocal acoustic measures of flat affect and alogia. J. Psychiatr. Res. 36 (5), 347–353. that expressive abnormalities reflect a domain criteria common American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental across severe mental illness diagnoses. It is interesting to note that, Disorders (DSM-IV), 4th ed. American Psychiatric Press, Washington, DC. Andreasen, N.C., 1984. The Scale for the Assessment of Negative Symptoms (SANS). The in some cases, clinical symptoms were associated with speech charac- University of Iowa, Iowa City, IA. teristics. Of note, negative and psychosis symptoms were associated Boersma, P., Weenink, D., 2006. Praat: Doing Phonetics by Computer (Version 4.4.05). with reduced speech production. It is not surprising that negative Cohen, A.S., Hong, S.L., 2011. Understanding constricted affect in schizotypy through computerized prosodic analysis. J. Pers. Disord. 25 (4), 478–491. symptoms were associated with at least some expressive characteris- Cohen, A.S., Alpert, M., Nienow, T.M., Dinzeo, T.J., Docherty, N.M., 2008. Computerized mea- tics because they are defined, at least in part, by blunted affect and surement of negative symptoms in schizophrenia. J. Psychiatr. Res. 42 (10), 827–836. psychomotor retardation. However, the finding that reduced speech Cohen, A.S., Minor, K.S., Najolia, G.M., Lee Hong, S., 2009. A laboratory-based procedure for measuring emotional expression from natural speech. Behav. Res. Meth. 41 (1), production was associated with psychosis (as opposed to schizophre- 204–212. nia specifically) is interesting in that psychosis is a symptom set com- Cohen, A.S., Hong, S.L., Guevara, A., 2010. Understanding emotional expression using mon to schizophrenia, depression and Bipolar I diagnoses. In the prosodic analysis of natural speech: refining the methodology. J. Behav. Ther. Exp. Psychiatry 41 (2), 150–157. present study, psychotic symptoms were also associated with higher Cohen, A.S., Kim, Y., Najolia, G.M., under review. Evaluating the “primacy” of dimin- levels of inflection, suggesting that only certain channels of expres- ished expressivity in schizophrenia and mood disorders. sion are reduced as a function of psychotic symptoms. Moving for- Collins, L.M., Blanchard, J.J., Biondo, K.M., 2005. Behavioral signs of schizoidia and – – ward, investigating why certain aspects of expressive characteristics schizotypy in social anhedonics. Schizophr. Res. 78 (2 3), 309 322. Fenton, W.S., McGlashan, T.H., 1991. Natural history of schizophrenia subtypes. II. Positive occur as a function of psychosis whereas others don't seems an im- and negative symptoms and long-term course. Arch. Gen. Psychiatry 48 (11), portant direction for future research. 978–986. This project has some limitations warranting mention. First, while First, M.B.G., Spitzer, R.L., Williams, J.B., 1996. User's Guide for the Structured Clinical Interview for DSM-IV Axis I Disorders — Research Version (SCID-I) Version 2.0, all of the subjects in this study were medicated and psychiatrically sta- February 1996 Final Version. Biometrics Research Department, New York State ble, it is beyond this study to meaningfully control for differences in Psychiatric Institute, New York. medication type, dosage or side effects; particularly with the modest Fleiss, J.L., 1986. The Design and Analysis of Clinical Experiments. Wiley, New York. Gur, R.E., Kohler, C.G., Ragland, J.D., Siegel, S.J., Lesko, K., Bilker, W.B., et al., 2006. Flat sample sizes employed. It is possible that sedation effects are responsi- affect in schizophrenia: relation to emotion processing and neurocognitive mea- ble for the null findings between patients with schizophrenia and those sures. Schizophr. Bull. 32 (2), 279–287. with mood disorders, or for the statistically significant relationships be- Herbener, E.S., Harrow, M., 2004. Are negative symptoms associated with functioning def- icits in both schizophrenia and nonschizophrenia patients? A 10-year longitudinal tween negative symptoms and expressive characteristics. Second, there analysis. Schizophr. Bull. 30 (4), 813–825. was no healthy control group for reference. The present findings are still Insel, T., Cuthbert, B., Garvey, M., Heinssen, R., Pine, D.S., Quinn, K., Sanislow, C., Wang, quite informative, particularly since the most interesting findings P., 2010. Research domain criteria (RDoC): toward a new classification framework for research on mental disorders. Am. J. Psychiatry 167 (7), 748–751. concern within-group differences. Regardless, future studies should Kulhara, P., Chadda, R., 1987. A study of negative symptoms in schizophrenia and de- include a non-psychiatric control group so absolute values of speech pression. Compr. Psychiatry 28 (3), 229–235. characteristics can be properly evaluated. Third, the modest sample Lader, M., 2000. Rating scales in schizophrenia: a review of their usefulness for size may have been underpowered for detecting small effect size differ- assessing atypical antipsychotics. CNS Drugs 14 (1), 23. Lang, P.J., Bradley, M.M., Cuthbert, B.N., 2005. International affective picture system ences between the schizophrenia and mood disorder groups. Finally, (IAPS): affective ratings of pictures and instruction manual. Technical Report A-6. we made no efforts to control for multiple analyses. It is unlikely that University of Florida, Gainesville, FL. this affected the results in any meaningful way, given that most of the Lukoff, D., Nuechterlein, K.H., Ventura, J., 1986. Manual for the expanded Brief Psychi- atric Rating Scale (BPRS). Schizophr. Bull. 12, 594–602. chief analyses yielded null findings. Meehl, P., 1962. Schizotaxia, schizotypy, schizophrenia. Am. Psychol. 17, 827–838. In closing, it is worth noting that the present findings provide further Mueser, K.T., Sayers, S.L., Schooler, N.R., Mance, R.M., Haas, G.L., 1994. A multisite inves- support for the use of acoustic analysis to understand expressive abnor- tigation of the reliability of the Scale for the Assessment of Negative Symptoms. Am. J. Psychiatry 151 (10), 1453–1462. malities in psychopathology. Though a modest literature employing Mueser, K.T., Pratt, S.I., Bartels, S.J., Forester, B., Wolfe, R., Cather, C., 2010. acoustic analysis in schizophrenia exists, few of these studies use this Neurocognition and social skill in older persons with schizophrenia and major technology to clarify their underpinnings. This is a knowledge gap that mood disorders: an analysis of gender and diagnosis effects. J. Neurolinguist 23 (3), 297–317. we are trying to redress. Acoustic analysis of natural speech appears to Parker, G., Hadzi-Pavlovic, D., Brodaty, H., Boyce, P., Mitchell, P., Wilhelm, K., et al., be a promising method of assessing and understanding expressive 1992. Predicting the course of melancholic and nonmelancholic depression. A nat- deficits across a wide array of pathology. uralistic comparison study. J. Nerv. Ment. Dis. 180 (11), 693–702. Tremeau, F., Malaspina, D., Duval, F., Correa, H., Hager-Budny, M., Coin-Bariou, L., et al., 2005. Facial expressiveness in patients with schizophrenia compared to depressed patients and nonpatient comparison subjects. Am. J. Psychiatry 162 (1), 92–101. Role of funding source Ventura, J., Nuechterlein, K.H., Subotnik, K.L., Gutkind, D., Gilbert, E.A., 2000. Symptom Funding for this study was provided by a Louisiana Board of Regents and National dimensions in recent-onset schizophrenia and mania: a principal components Institute of Mental Health (R03 MH092622) grant to the primary author. The funding analysis of the 24-item Brief Psychiatric Rating Scale. Psychiatry Res. 97 (2–3), agencies had no further role in the study design; in the collection, analysis and inter- 129–135. pretation of data; in the writing of the report; and in the decision to submit the Wilkinson, G.S., Robertson, G.J., 2006. Wide Range Achievement Test 4 (WRAT4). Psycho- paper for publication. logical Assessment Resources, Inc.