Schizophrenia Research 152 (2014) 235–241

Contents lists available at ScienceDirect

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

Abnormalities in the processing of emotional prosody from single words

in schizophrenia
Ana P. Pinheiro a,b, Neguine Rezaii b, Andréia Rauber c, Taosheng Liu d, Paul G. Nestor e,
Robert W. McCarley b, Óscar F. Gonçalves a, Margaret A. Niznikiewicz b
a
Neuropsychophysiology Lab, CIPsi, School of Psychology, University of Minho, Braga, Portugal
b
Clinical Neuroscience Division, Laboratory of Neuroscience, Department of Psychiatry, Boston VA Healthcare System, Brockton Division and Harvard Medical School, Brockton, MA, United States
c
Catholic University of Pelotas, Pelotas, Brazil
d
Department of Psychology, Second Military Medical University (SMMU), Shanghai, China
e
University of Massachusetts, Boston, MA, United States

a r t i c l e i n f o a b s t r a c t

Article history: Background: Abnormalities in emotional prosody processing have been consistently reported in schizophrenia
Received 4 July 2013 and are related to poor social outcomes. However, the role of stimulus complexity in abnormal emotional prosody
Received in revised form 25 October 2013 processing is still unclear.
Accepted 28 October 2013 Method: We recorded event-related potentials in 16 patients with chronic schizophrenia and 16 healthy controls
Available online 14 December 2013
to investigate: 1) the temporal course of emotional prosody processing; and 2) the relative contribution of
prosodic and semantic cues in emotional prosody processing. Stimuli were prosodic single words presented in
Keywords:
Schizophrenia
two conditions: with intelligible (semantic content condition—SCC) and unintelligible semantic content (pure
Emotional prosody prosody condition—PPC).
Language Results: Relative to healthy controls, schizophrenia patients showed reduced P50 for happy PPC words, and
Event-related potentials reduced N100 for both neutral and emotional SCC words and for neutral PPC stimuli. Also, increased P200 was
observed in schizophrenia for happy prosody in SCC only. Behavioral results revealed higher error rates in schizo-
phrenia for angry prosody in SCC and for happy prosody in PPC.
Conclusions: Together, these data further demonstrate the interactions between abnormal sensory processes
and higher-order processes in bringing about emotional prosody processing dysfunction in schizophrenia. They
further suggest that impaired emotional prosody processing is dependent on stimulus complexity.
© 2013 Published by Elsevier B.V.

1. Introduction In healthy subjects, perception of emotional prosody is thought to

reflect three interacting stages: 1) sensory processing of a speech signal;
Among the most significant predictors of long-term disability in 2) implicit categorization of salient acoustic features into emotional and
schizophrenia (e.g., Couture et al., 2006) is impaired detection and non-emotional features; and 3) explicit evaluation and assignment of
recognition of emotions from voice, i.e., emotional prosody [EP]. Affect emotional meaning to a speech signal (Schirmer and Kotz, 2006;
recognition from both voice and face is an aspect of social cognition, Paulmann and Kotz, 2008; Paulmann et al., 2010). Event-related poten-
which has been recently recognized as an important predictor of func- tial (ERP) studies demonstrated that the first two stages are indexed by
tional outcomes at all stages of schizophrenia pathology: clinical high N100 and P200, respectively (Paulmann and Kotz, 2008; Paulmann
risk (Addington et al., 2008; Green et al., 2012), first episode (Horan et al., 2010; Pinheiro et al., 2012).
et al., 2012) and chronic schizophrenia (Kee et al., 2003; Kucharska- Despite the importance of a detailed understanding of emotional
Pietura et al., 2005; Green et al., 2012). While face processing abnormal- prosody processing deficits in schizophrenia, few studies have exam-
ity in schizophrenia has been well characterized (e.g., Li et al., 2010), ined these abnormalities and their underlying neural mechanisms
voice and prosody processing have been understudied, especially are not well understood. Recent studies suggested that sensory-based
using event-related potential (ERP) approaches, which remain the dysfunction might not exclusively account for abnormal prosody pro-
only tool to examine temporal changes in neurophysiological events cessing in schizophrenia. Instead, an interaction between dysfunctional
that correspond to early stages of analysis of a speech signal. The sensory and higher-order cognitive processes may better explain it
existing studies on vocal emotional processing include just a handful (Leitman et al., 2010, 2011; Pinheiro et al., 2012). A recent ERP study
of behavioral (e.g., Edwards et al., 2001), functional magnetic resonance provided further evidence for these abnormalities (Pinheiro et al.,
imaging (fMRI—e.g., Mitchell et al., 2004; Leitman et al., 2011) and ERP 2012). This study investigated prosody processing in 15 chronic schizo-
investigations (Pinheiro et al., 2012). phrenia patients and 15 healthy controls (HC). Additionally, it explored

0920-9964/$ – see front matter © 2013 Published by Elsevier B.V.

http://dx.doi.org/10.1016/j.schres.2013.10.042
236 A.P. Pinheiro et al. / Schizophrenia Research 152 (2014) 235–241

the relative contributions of prosodic and semantic cues. Stimuli were Table 1
prosodic sentences with intelligible (semantic content condition—SCC) Demographic and clinical characteristics of participants.

and unintelligible semantic content (pure prosody condition—PPC). Variable Healthy Schizophrenia patients p valuea
The ERP effects occurred within the first 200 ms from the sentence controls (n = 16)
onset in both groups (Pinheiro et al., 2012), supporting previous studies' (n = 17)

results (Paulmann and Kotz, 2008; Paulmann et al., 2010). The results Age (years) 48.13 ± 5.66 48.86 ± 7.40 .750
revealed abnormalities in the three stages of prosody processing Women, n 7 5
Education (years) 15.18 ± 1.64 14.00 ± 2.42 .119
in schizophrenia, which were more pronounced for prosodic SCC
Subject's SESb 2.13 ± 0.81 2.93 ± 1.14 .033⁎
sentences. Less negative N100 suggested abnormal sensory processing Parental SES 2.44 ± 0.81 2.79 ± 1.53 .434
of prosodic SCC sentences irrespective of valence. Increased P200 to Handednessc 0.81 ± 0.15 0.79 ± 0.21 .848
angry and happy prosodic stimuli in the SCC, and to happy stimuli in
Neurocognitive data
the PPC suggested abnormal detection of emotional salience. Behavioral Full scale composite 99.33 ± 12.30 92.79 ± 14.32 .227
results revealed impaired cognitive evaluation of the emotional signifi- score
cance of angry SCC and neutral PPC sentences. Verbal comprehension 99.08 ± 11.47 95.93 ± 15.82 .572
In view of a critical need for a systematic study of emotional prosody composite score
Working memory 105.33 ± 14.22 92.86 ± 12.90 .049⁎
processing in schizophrenia, the current study extended our previous
composite score
work, by investigating the temporal course of prosody processing Processing speed 101.17 ± 89.64 89.64 ± 14.87 .107
using single words with both intelligible (SCC) and unintelligible seman- composite score
tic content (PPC). Based on language studies demonstrating differences
Clinical data
in the processing of words in a sentence vs. in isolation (e.g., Van Onset age (years) NA 30.07 ± 11.23 NA
Petten, 1995) and effects of phrasal length and complexity on prosodic Duration (years) NA 19.47 ± 10.95 NA
processing (Wheeldon and Lahiri, 1997; Krivokapi, 2007), we reasoned Chlorpromazine EQ NA 356.78 ± 294.56 NA
(mg)
that prosody processing of sentences may differ from that of single
Antipsychotic NA Typical (fluphenazine NA
words. For example, the processing of words embedded in a sentence medication type decanoate, proloxin decanoate,
is susceptible to syntactic and semantic constraints imposed by a sen- haloperidol) = 3;
tence context, which can modify many aspects of their processing Atypical (risperidone,
(e.g., Van Petten, 1995). Furthermore, in relation to words in isolation, olanzapine, ziprasidone,
quetiapine, aripiprazole) = 11
the processing of a sentence demands more working memory and
Other psychotropic NA Antidepressants (sertraline, NA
attention resources, as meaning is built up across the course of the medication citalopram, bupoprion,
sentence (e.g., Van Petten, 1995). Thus, considering the attentional trazodone) = 4
(Nestor et al., 2001; Laurens et al., 2005) and verbal working memory Benzodiazepines (lorazepam,
clonazepam) = 4
deficits (Menon et al., 2001; Silver et al., 2003) often reported in
Lithium carbonate = 2
schizophrenia, the processing of prosodic information may be more Valproic acid = 3
impaired in sentences than in single words. PANSS delusions NA 4.88 ± 2.16 NA
Because of its excellent temporal resolution, we used ERPs to PANSS conceptual NA 2.50 ± 1.10 NA
address the role of stimulus complexity in the first two stages of emo- disorganization
PANSS hallucinations NA 4.00 ± 2.19 NA
tional prosody processing: the sensory processing of prosodic informa-
PANSS positive scale NA 20.25 ± 8.19 NA
tion (N100) and the detection of its emotional salience (P200), both PANSS negative scale NA 22.88 ± 9.76 NA
processes not accessible to behavioral probes. We also collected data PANSS general NA 38.56 ± 11.70 NA
on accuracy of prosody recognition to shed light on a later stage of emo- psychopathology
tional prosody processing, i.e. the assignment of emotional meaning to PANSS total NA 81.69 ± 25.92 NA
psychopathology
a voice signal. We hypothesized that if impaired prosody processing is SANS total NA 10.59 ± 5.44 NA
not dependent on stimulus complexity, similar abnormalities to those SAPS total NA 9.63 ± 3.05 NA
reported in Pinheiro et al. (2012) will be observed in the current
Notes. All values represent mean ± SD. SES = socioeconomic status; Chlorpromazine
study. However, if stimulus complexity matters, we expected less EQ = chlorpromazine equivalent dose; NA = not applicable.
severe prosody processing abnormalities in the single word relative to a
Independent samples t-test tested for group differences in sociodemographic and
the sentence prosody processing study. neurocognitive measures.
b
Hollingshead Four-Factor Index of Social Status (Hollingshead, 1976).
Considering previous studies demonstrating an association between c
Edinburgh Handedness Inventory (Oldfield, 1971).
deficits in emotional prosody recognition and positive symptomatology ⁎ p b 0.05.
(Poole et al., 2000; Rossell and Boundy, 2005; Shea et al., 2007), and be-
tween increased P200 amplitude for happy prosody and delusions
(Pinheiro et al., 2012), we predicted that ERP abnormalities amplitude
would be associated with positive symptomatology scores. DSM-IV diagnosis of drug or alcohol abuse (APA, 2000) in the last year
prior to EEG assessment; full scale intelligence quotient (IQ) above 85
2. Method (Wechsler, 2008); no hearing, vision or upper body impairment. For
HC, additional inclusion criteria were: no history of Axis I–II disorders
2.1. Participants (First et al., 1995, 2002); no history of Axis I disorder in first or
second-degree relatives (Andreasen et al., 1977).
Sixteen patients with a diagnosis of chronic schizophrenia and 17 Patients were diagnosed (screened for HC) using the SCID-I and
HC matched for age, handedness and parental socioeconomic status SCID-II (First et al., 1995, 2002). Symptom severity was assessed with
(Hollingshead, 1976) participated in this study (Table 1). Subjects had the Positive and Negative Syndrome Scale (PANSS-Kay et al., 1987),
normal hearing as assessed by audiometry, and normal or corrected to the Scale for the Assessment of Negative Symptoms (SANS-Andreasen,
normal vision. Patients were recruited at the Veterans Affairs Hospital, 1983) and the Scale for the Assessment of Positive Symptoms (SAPS-
Brockton and HC were recruited from Internet advertisements. Andreasen, 1984) (Table 1).
The inclusion criteria were: English as first language; right handed- All participants had the procedures fully explained to them and read
ness (Oldfield, 1971); no history of neurological illness; no history of and signed an informed consent form.
 A.P. Pinheiro et al. / Schizophrenia Research 152 (2014) 235–241 237

2.2. Stimuli

Stimuli used in the SCC were 40 words with neutral semantic con-
tent and short length (e.g., “card”, “pen”—see Supplementary Material).
Words were controlled for frequency (M = 10.38 ± 11.05), familiarity
(M = 582.37 ± 25.88), age of acquisition (M = 232.54 ± 55.15), con-
creteness (M = 594.60 ± 43.79), and number of letters (M = 5.14 ±
1.81), based on the MRC Psycholinguistic Database (Coltheart, 1981;
Wilson, 1998).
Words were recorded by an American English native speaker with
training in theater techniques, with neutral and emotional prosody
(happy; angry) using an Edirol R-09 recorder and a CS-15 cardioid-
type stereo microphone, at a sampling rate of 22 kHz and 16-bit quan-
tization. Words' pitch, intensity and duration were compared across
conditions (Table 2). Duration of happy words was longer than duration Fig. 1. Illustration of an experimental trial.
of angry (p b 0.01) and neutral words (p b 0.001). Mean pitch was
higher for happy relative to both angry (p b 0.001) and neutral words
(p b 0.001), and for angry relative to neutral words (p b 0.001). Mean was counterbalanced across subjects. Each response key was marked
intensity did not differ across emotion types (p N 0.05). with an emoticon to minimize working memory demands. A 2000 ms
Fifteen subjects (7 female) who did not participate in the ERP sessions inter-stimulus interval separated the end of an event and the beginning
assessed the valence of words' intonation. Angry words were rated as of the next one. A short pause was provided after 15 words. During the
“angry” by 96.11%, happy words were rated as “happy” by 99.18%, and experiment, no feedback was provided.
neutral words were rated as “neutral” by 96.66% of participants.
In PPC, the same stimuli were distorted to make their semantic con- 2.4. Data acquisition and analysis
tent unintelligible by using Praat software (Boersma and Weenink,
2006). The phones of each SCC word were manually replaced by phones EEG was recorded with custom-made electrode caps with a 64-
produced by the same speaker, preserving both the original voice and channel BioSemi Active-Two system (BioSemi B.V., The Netherlands).
prosodic features. All fricatives were replaced with the phone [s], all It was acquired in a continuous mode at a digitization rate of 512 Hz,
stop consonants with [t], all glides with [j], all stressed vowels with with a bandpass of 0.01–100 Hz. Blinks and eye movements were
[æ], and all unstressed vowels with [Ə]. monitored through electrodes placed on the left and right temples and
below the left eye.
2.3. Procedure EEG data were processed offline using Brain Analyzer 2 package
(Brain Products, Germany), and re-referenced offline to the mathe-
Each participant was seated comfortably at a distance of 100 cm matical average of the left and right mastoids. Individual ERP epochs
from a computer monitor in a sound-attenuating chamber. The experi- were created for each prosody type (neutral, happy, angry) in each
mental session was divided into two blocks (block1: SCC words; block2: word condition (SCC, PPC), with −200 ms pre-stimulus baseline and
PPC words). Block order was counterbalanced. Each block contained 900 ms post-stimulus epoch. Eye blinks and movement artifacts were
105 words of different prosody types (35 neutral, 35 happy, 35 angry). corrected by the Gratton et al. (1983) method. EEG epochs containing
The remaining five words of each valence and type (SCC, PPC), muscle activity or amplifier blocking were rejected offline before
from the original list of 40 words, were presented in the practice averaging (+/−100 μV criterion). After artifact rejection, at least 75%
block. Stimuli were presented binaurally through headphones at a com- of trials per condition per subject entered the analyses. The number of
fortable sound level. Superlab Pro software package (2008) controlled individual trials did not differ between groups (p N .05).
stimulus delivery. The inspection of grand average waveforms (Figs. 2 and 3) revealed
Before each experimental block, participants were given a brief three main components with predominantly central distribution: a
training with feedback. Fig. 1 illustrates an experimental trial. Before positivity occurring around 50 ms (P50), a negativity occurring around
each word onset, a fixation cross was presented centrally on the screen 100 ms (N100), and a positivity occurring around 200 ms (P200).
for 1000 ms, and was kept during word presentation to minimize eye Temporal windows were then selected for P50, N100 and P200
movements. After 1000 ms, a question mark signaled the beginning of based on the visual inspection of the waveforms. Mean amplitude was
the response time (5 s). Subjects were instructed to make a decision calculated between 30 and 125 ms (P50), 125 and 190 ms (N100),
whether a word was spoken with a neutral, happy, or angry tone of and 220 and 320 ms (P200), post-stimulus onset, at central electrodes
voice by pressing one of the three buttons. The order of button presses (Cz, C3, C4).

Table 2
Acoustic properties of words with angry, happy and neutral prosody in the semantic content (SCC) and pure prosody (PPC) conditions.

Semantic status Emotion Acoustic properties

Duration F0 Intensity

Min Mean Max Min Mean Max

SCC Angry 639.00 (19.45) 163.09 (2.13) 271.94 (4.12) 368.15 (5.68) 45.67 (1.05) 78.64 (0.39) 86.24 (0.28)
Happy 718.21 (19.89) 177.58 (11.44) 371.03 (12.45) 574.94 (13.53) 50.58 (0.84) 79.27 (0.30) 86.49 (0.26)
Neutral 663.23 (22.12) 156.91 (2.04) 183.70 (1.24) 228.09 (4.36) 49.82 (1.41) 78.97 (0.28) 85.09 (0.22)
PPC Angry 653.94 (25.65) 170.55 (5.15) 265.00 (4.38) 363.09 (6.98) 43.18 (0.83) 77.85 (0.36) 84.88 (0.39)
Happy 722.27 (24.34) 173.03 (8.76) 377.73 (10.01) 559.00 (12.91) 48.27 (1.05) 78.73 (0.28) 86.00 (0.26)
Neutral 629.24 (28.83) 156.70 (1.59) 181.55 (1.52) 219.09 (2.55) 48.39 (1.68) 77.82 (0.35) 83.79 (0.22)

Notes. Mean (standard error); min = minimum; max = maximum.

238 A.P. Pinheiro et al. / Schizophrenia Research 152 (2014) 235–241

Fig. 2. Grand average waveforms for neutral, happy and angry prosody in the semantic content condition (SCC) and pure prosody condition (PPC) at Cz, in healthy controls (HC) and
schizophrenia patients (SZ).

Fig. 3. Grand average waveforms showing group contrasts for neutral, happy, and angry prosody in the semantic content condition (SCC) and pure prosody condition (PPC) at Cz.
 A.P. Pinheiro et al. / Schizophrenia Research 152 (2014) 235–241 239

2.5. Statistical analyses p = 0.020) indicated more incorrect responses for angry SCC words
(p = 0.036) and happy PPC words (p = 0.029) in schizophrenia
For the statistical analysis, the PAWS 20.00 (SPSS Inc., USA) software (Fig. 4). However, no main effect of group was observed (p N 0.05).
package was used. Only significant results are presented (p b 0.05).
3.3. Correlational analyses
2.5.1. ERP data
Repeated measures analyses of variance (ANOVAs) were computed Two-tailed Spearman's rho correlation analyses were conducted in
for the between-group comparisons of N100 and P200 peak amplitude, an exploratory analysis of the relationship between schizophrenia ab-
with semantic status (SCC, PPC), emotion (neutral, happy, angry), normalities in P50 (happy PPC), N100 (neutral, happy, and angry SCC;
and electrodes (Cz, C3, C4) as within-subject factors and group as a neutral PPC) and P200 (happy SCC) amplitude at Cz and: 1) clinical
between-subject factor, using IBM SPSS Statistics 20 (SPSS Inc., USA). symptoms (PANSS), medication (chlorpromazine equivalent) and ill-
ness duration; 2) neurocognitive data (WAIS composite scores); and
2.5.2. Accuracy data 3) behavioral indices of prosody recognition. The significance level
A repeated measures ANOVA with semantic status and emotion as was adjusted for multiple comparisons using Bonferroni correction.
within-subjects factors and group as between-subjects factor tested No significant correlations were found (p N 0.05).
group differences in behavioral accuracy.
Analyses were corrected for non-sphericity using the Greenhouse– 4. Discussion
Geisser method (the original df is reported). All significance levels are
two-tailed with the preset significance alpha level of p b 0.05. Main This study extended and clarified our previous findings for prosodic
effects were followed with pairwise comparisons between conditions, sentences (Pinheiro et al., 2012). ERP and behavioral findings showed
using the Bonferroni adjustment for multiple comparisons. group differences that spanned the three stages of prosody processing
and interacted with the semantic status of words. ERP effects were
3. Results observed within the first 200 ms. In addition to N100 and P200, we ob-
served prosodic effects in an earlier time window around 50 ms (P50),
3.1. ERP data (Figs. 2 and 3) corroborating the sensitivity of P50, N100 and P200 components to
prosodic manipulations in speech sounds (Paulmann and Kotz, 2008;
3.1.1. P50 amplitude Paulmann et al., 2010; Liu et al., 2012; Pinheiro et al., 2012). Schizophre-
A significant group × semantic status × emotion interaction was nia patients showed a markedly different P50, N100 and P200 pattern as
observed (F(2, 62) = 6.603, p b 0.01). We followed-up this interaction a function of both semantic status and emotion type relative to HC.
with subsequent ANOVAs for each semantic status condition separately. Reduced P50 amplitude for happy PPC prosody was observed in
A significant group × emotion interaction was observed in the PPC only schizophrenia relative to HC. P50 has been reported in studies of audito-
(F(2, 62) = 4.872, p = 0.016): groups differed in the processing of ry gating (e.g., Boutros et al., 2004) and has been considered an index of
happy PPC prosody (p b 0.01), with reduced P50 amplitude in patients the formation of sensory memory traces at the level of the primary au-
relative to HC. ditory cortex (Haenschel et al., 2005). The existing evidence suggests
that P50 amplitude may be modulated by the physical properties of
3.1.2. N100 amplitude the eliciting stimulus (Chen et al., 1997; Ninomiya et al., 2000) and by
A main effect of emotion (F(2, 62) = 6.723, p b 0.01) revealed attention (Erwin et al., 1998). Also, in our previous study (Liu et al.,
that N100 was more negative for angry relative to neutral prosody 2012) with non-verbal vocalizations, emotion effects were found at
(p b 0.01) and tended to be more negative for angry relative to happy the level of P50 (Liu et al., 2012). Reduced P50 amplitude has been con-
prosody (p = 0.083) in both groups. A significant group × semantic sistently demonstrated in schizophrenia (e.g., Potter et al., 2006). In our
status × emotion interaction (F(2, 62) = 4.638, p = 0.02) indicated study, reduced P50 for happy PPC in schizophrenia points to abnormal
differences in the way patients and HC processed prosodic stimuli at early somatosensory information processing that is stimulus specific.
the sensory level.
Follow-up separate repeated-measures ANOVAs for each semantic
status condition showed a significant group effect for the SCC (F(1,
31) = 8.395, p b 0.01): N100 was overall less negative in the schizo-
phrenia group relative to HC. In addition, a significant group × emotion
interaction was observed in the PPC (F(2, 62) = 3.874, p = 0.027).
Pairwise comparisons indicated less negative N100 in schizophrenia
relative to HC subjects in the neutral condition only (p = 0.039).

3.1.3. P200 amplitude

A significant group × semantic status × emotion interaction (F(2,
62) = 5.476, p b 0.01) indicated differences in the way groups inte-
grated acoustic information into an emotional percept. Separate
ANOVAs were subsequently computed for each semantic status condi-
tion. A significant group × emotion interaction was observed for SCC
(F(2, 62) = 3.215, p = 0.049). Subsequent pairwise comparisons indi-
cated more positive P200 for happy prosody in patients relative to HC
(p = 0.01). No significant effects were observed for PPC.

3.2. Accuracy data

More correct responses were found in SCC relative to PPC (main Fig. 4. Percentage of correct responses in the recognition of emotional prosody in both
effect of semantic status—F(1, 31) = 45.606, p b 0.001). A significant semantic content condition (SCC) and pure prosody condition (PPC) in healthy controls
group × semantic status × emotion interaction (F(2, 62) = 4.327, (HC) and schizophrenia patients (SZ).
240 A.P. Pinheiro et al. / Schizophrenia Research 152 (2014) 235–241

Reduced N100 amplitude in schizophrenia was found to both emo- not identical. They additionally suggest that stimulus complexity may
tional and neutral SCC stimuli, as well as to neutral PPC stimuli. The differently impact sensory and early categorization stages of prosody
N100 component is related to early auditory encoding, and its ampli- processing.
tude is modulated by the physical properties of the stimuli and by allo- Furthermore, the overall reduced emotional recognition accuracy
cation of attentional resources (Rosburg et al., 2008). P50 and N1 are observed in the sentences study contrasted with specific deficits in the
thought to represent distinct aspects of information processing recognition of angry SCC prosody and happy PPC prosody in single
(Boutros et al., 2004). Considering the functional role of N100 as an words. Since prosody processing relies on the continuous monitoring
index of initial sensory processing of the prosodic signal (Schirmer of dynamically changing acoustic cues underlying an emotional tone,
and Kotz, 2006), these findings support deficits in sensory processing greater working memory and attention demands exist for sentences
of vocal information (e.g., Pinheiro et al., 2012) that were enhanced vs. single words. Accordingly, they were associated with more errors
when semantic information was present. Given that N100 generators in identifying sentence relative to single word prosodic stimuli.
are located mainly in supratemporal plane and superior temporal
gyrus (Naatanen and Picton, 1987), reduced N100 amplitude may re- 4.1. Limitations
flect functional and structural brain changes in temporal structures Limitations of this study are a sample composed by medicated
that are a central feature of the schizophrenia diagnosis (e.g., Shenton chronic schizophrenia patients. Future research with unmedicated and
et al., 1992). However, since N100 cannot be directly related to a single first-episode patients will overcome some of the limitations associated
cortical process and is influenced by many individual-related variables, with medication and chronicity.
we cannot rule out the contributions of other factors, such as attention
or arousal. Role of funding source
This work was supported by two grants awarded to A.P.P.: Post-Doctoral Grant no.
Specific abnormalities were noted in the second stage of prosody
SFRH/BD/35882/2007 funded by Fundação para a Ciência e a Tecnologia (FCT, Portugal);
processing as indexed by increased P200 amplitude to happy SCC and research grant no. PTDC/PSI-PCL/116626/2010, funded by FCT and FEDER (Fundo
words only in schizophrenia relative to HC. The P200 is primarily gen- Europeu de Desenvolvimento Regional) through the European programs QREN (Quadro de
erated in the temporal cortex (such as the planum temporale and the Referência Estratégico Nacional) and COMPETE (Programa Operacional Factores de
auditory association complex, area 22—Godey et al., 2001), even Competitividade). It was additionally supported by two grants from the National Institute
of Mental Health (no. RO1 MH 040799 awarded to R.W.M. and no. RO3 MH 078036
though frontal areas are also involved (McCarley et al., 1991). Consid-
awarded to M.A.N.).
ering the role of P200 as an index of the emotional salience of a vocal
stimulus (Paulmann and Kotz, 2008), as proposed in the multi-stage Contributors
model of emotional prosody processing (Schirmer and Kotz, 2006), in- Ana P. Pinheiro and Margaret Niznikiewicz designed the study and wrote the protocol.
creased P200 for happy prosody might indicate a specific impairment Ana P. Pinheiro, Margaret Niznikiewicz, Neguine Rezaii and Taosheng Liu collected and
analyzed the data. All authors collaborated in the statistical analysis. Andréia Rauber
in categorizing happy auditory emotional percepts as “salient”. How-
edited the stimuli and completed the acoustic analyses of the stimuli. Paul G. Nestor did
ever, this was the case only when happy prosodic information was the clinical and neuropsychological testing of participants. Ana P. Pinheiro and Margaret
embedded in intelligible speech suggesting that sensory cues were Niznikiewicz wrote the first draft of the manuscript. All authors contributed to and have
used differently in the two conditions. Additionally, given the sensitivity approved the final manuscript.
of P200 to task difficulty (increased P200 amplitude related to increased
cognitive effort—Lenz et al., 2007), it is plausible that the salience of Conflict of interest
All authors report no competing interests.
positive social information was more difficult to extract for schizophre-
nia patients. Also, given that all stimuli had neutral semantic content
Acknowledgments
but could carry emotional intonation, we cannot rule out the effects We are grateful to all the participants of this study for their contribution to science.
of incongruity (semantic vs. prosodic) on P200 amplitude (Scholten We are also grateful to Elizabeth Thompson and Israel Molina for their help with data ac-
et al., 2008). quisition, and to Victoria Choate for her invaluable help with voice recording.
Two major conclusions arise from P50, N100 and P200 findings in
the current study: 1) the fact that group differences were not observed Appendix A. Supplementary data
for all types of prosodic stimuli speaks against a generalized prosodic
impairment and suggests that prosodic abnormalities may be depen- Supplementary data to this article can be found online at http://dx.
dent on stimulus type; 2) abnormalities in the processing of emotional doi.org/10.1016/j.schres.2013.10.042.
but not neutral cues seem to be more pronounced when speech's
semantic content is intelligible, suggesting that abnormalities in the References
processing of both semantic and prosodic aspects of voice interact
Addington, J., Penn, D., Woods, S.W., Addington, D., Perkins, D.O., 2008. Facial affect recog-
since early stages of prosody processing. nition in individuals at clinical high risk for psychosis. Br. J. Psychiatry 192 (1), 67–68.
Behavioral data, indexing the integration of emotionally significant American Psychiatric Association, 2000. Diagnostic and Statistical Manual of Mental
acoustic cues (Schirmer and Kotz, 2006), indicated that emotional Disorders: DSM-IV-TR. American Psychiatric Association Press, Washington, DC.
Andreasen, N.C., 1983. Scale for Assessment of Negative Symptoms (SANS). University of
prosody recognition was better in SCC relative to PPC in both groups, Iowa, Iowa City, IA.
confirming our initial hypothesis (and also Pinheiro et al., 2012). Andreasen, N.C., 1984. Scale for Assessment of Positive Symptoms (SAPS). University of
Given the absence of a memory representation for unintelligible stimuli Iowa, Iowa City, IA.
Andreasen, N.C., Endicott, J., Spitzer, R.L., Winokur, G., 1977. The family history method
to facilitate predictive processes, this result likely reflects increased task using diagnostic criteria: reliability and validity. Arch. Gen. Psychiatry 34, 1229–1235.
demands. Additionally, schizophrenia patients made more errors in Boersma, P., Weenink, D., 2006. Praat: Doing Phonetics by Computer (Version 5.0.43)
identifying emotional but not neutral prosody. This result suggests that [Computer Program].
Boutros, N.N., Korzyukov, O., Jansen, B., Feingold, A., Bell, M., 2004. Sensory gating deficits
deficits in recognizing emotional prosody in single words depend both
during the mid-latency phase of information processing in medicated schizophrenia
on emotion type and on semantic status. patients. Psychiatry Res. 126 (3), 203–215.
Finally, we note differences in both ERP and behavioral results re- Chen, C.H., Ninomiya, H., Onitsuka, T., 1997. Influence of reference electrodes, stimulation
characteristics and task paradigms on auditory P50. Psychiatry Clin. Neurosci. 51 (3),
ported in this and in our previous study using prosodic sentences
139–143.
(Pinheiro et al., 2012). In our previous study, P200 was increased to Coltheart, M., 1981. The MRC psycholinguistics database. Q. J. Exp. Psychol. A 33A,
happy SCC and PPC sentences and to angry SCC sentences; here, P200 497–505.
abnormalities were observed only for happy SCC words. These ERP Couture, S.M., Penn, D.L., Roberts, D.L., 2006. The functional significance of social cognition
in schizophrenia: a review. Schizophr. Bull. 32 (Suppl. 1), S44–S63.
differences suggest that while processing prosody in sentences and Edwards, J., Pattison, P.E., Jackson, H.J., Wales, R.J., 2001. Facial affect and affective prosody
words evokes similar ERP components, the processes involved are recognition in first-episode schizophrenia. Schizophr. Res. 48 (2–3), 235–253.
 A.P. Pinheiro et al. / Schizophrenia Research 152 (2014) 235–241 241

Erwin, R.J., Turetsky, B.I., Moberg, P., Gur, R.C., Gur, R.E., 1998. P50 abnormalities in Menon, V., Anagnoson, R.T., Mathalon, D.H., Glover, G.H., Pfefferbaum, A., 2001. Functional
schizophrenia: relationship to clinical and neuropsychological indices of attention. neuroanatomy of auditory working memory in schizophrenia: relation to positive
Schizophr. Res. 33 (3), 157–167. and negative symptoms. Neuroimage 13 (3), 433–446.
First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.M., 1995. Structured Clinical Interview Mitchell, R.L., Elliott, R., Barry, M., Cruttenden, A., Woodruff, P.W., 2004. Neural response
for DSM-IV Axis II Personality Disorders (SCID-II, Version 2.0). Biometrics Research to emotional prosody in schizophrenia and in bipolar affective disorder. Br.
Department, New York State Psychiatric Institute, New York, NY. J. Psychiatry 184, 223–230.
First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.W., 2002. Structured Clinical Interview Naatanen, R., Picton, T., 1987. The N1 wave of the human electric and magnetic response
for DSM-IV Axis I Diagnosis-Patient Edition (SCID-I/P, Version 2.0). Biometric to sound: a review and an analysis of the component structure. Psychophysiology
Research Department, New York State Psychiatric Institute, New York, NY. 24 (4), 375–425.
Godey, B., Schwartz, D., de Graaf, J.B., Chauvel, P., Liegeois-Chauvel, C., 2001. Neuromagnetic Nestor, P.G., Han, S.D., Niznikiewicz, M., Salisbury, D., Spencer, K., Shenton, M.E., McCarley,
source localization of auditory evoked fields and intracerebral evoked potentials: R.W., 2001. Semantic disturbance in schizophrenia and its relationship to the cogni-
a comparison of data in the same patients. Clin. Neurophysiol. 112 (10), 1850–1859. tive neuroscience of attention. Biol. Psychol. 57 (1–3), 23–46.
Gratton, G., Coles, M.G., Donchin, E., 1983. A new method for off-line removal of ocular Ninomiya, H., Sato, E., Onitsuka, T., Hayashida, T., Tashiro, N., 2000. Auditory P50 obtained
artifact. Electroencephalogr. Clin. Neurophysiol. 55 (4), 468–484. with a repetitive stimulus paradigm shows suppression to high-intensity tones.
Green, M.F., Bearden, C.E., Cannon, T.D., Fiske, A.P., Hellemann, G.S., Horan, W.P., Kee, K., Psychiatry Clin. Neurosci. 54 (4), 493–497.
Kern, R.S., Lee, J., Sergi, M.J., Subotnik, K.L., Sugar, C.A., Ventura, J., Yee, C.M., Oldfield, R.C., 1971. The assessment and analysis of handedness: the Edinburgh inventory.
Nuechterlein, K.H., 2012. Social cognition in schizophrenia, part 1: performance Neuropsychologia 9 (1), 97–113.
across phase of illness. Schizophr. Bull. 38 (4), 854–864. Paulmann, S., Kotz, S.A., 2008. Early emotional prosody perception based on different
Haenschel, C., Vernon, D.J., Dwivedi, P., Gruzelier, J.H., Baldeweg, T., 2005. Event-related speaker voices. Neuroreport 19 (2), 209–213.
brain potential correlates of human auditory sensory memory-trace formation. Paulmann, S., Seifert, S., Kotz, S.A., 2010. Orbito-frontal lesions cause impairment during
J. Neurosci. 25 (45), 10494–10501. late but not early emotional prosodic processing. Soc. Neurosci. 5 (1), 59–75.
Hollingshead, A., 1976. Four Factor Index of Social Status. Yale University Social Sciences Pinheiro, A.P., Del Re, E., Mezin, J., Nestor, P.G., Rauber, A., McCarley, R.W., Goncalves, O.F.,
Library, New Haven, CT. Niznikiewicz, M.A., 2012. Sensory-based and higher-order operations contribute to
Horan, W.P., Green, M.F., DeGroot, M., Fiske, A., Hellemann, G., Kee, K., Kern, R.S., Lee, J., abnormal emotional prosody processing in schizophrenia: an electrophysiological
Sergi, M.J., Subotnik, K.L., Sugar, C.A., Ventura, J., Nuechterlein, K.H., 2012. Social investigation. Psychol. Med. 1–16.
cognition in schizophrenia, part 2: 12-month stability and prediction of functional Poole, J.H., Tobias, F.C., Vinogradov, S., 2000. The functional relevance of affect recognition
outcome in first-episode patients. Schizophr. Bull. 38 (4), 865–872. errors in schizophrenia. J. Int. Neuropsychol. Soc. 6 (6), 649–658.
Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The Positive and Negative Syndrome Scale Potter, D., Summerfelt, A., Gold, J., Buchanan, R.W., 2006. Review of clinical correlates
(PANSS) for schizophrenia. Schizophr. Bull. 13 (2), 261–276. of P50 sensory gating abnormalities in patients with schizophrenia. Schizophr. Bull.
Kee, K.S., Green, M.F., Mintz, J., Brekke, J.S., 2003. Is emotion processing a predictor of 32 (4), 692–700.
functional outcome in schizophrenia? Schizophr. Bull. 29 (3), 487–497. Rosburg, T., Boutros, N.N., Ford, J.M., 2008. Reduced auditory evoked potential component
Krivokapi, J., 2007. Prosodic planning: effects of phrasal length and complexity on pause N100 in schizophrenia—a critical review. Psychiatry Res. 161 (3), 259–274.
duration. J. Phon. 35 (2), 162–179. Rossell, S.L., Boundy, C.L., 2005. Are auditory-verbal hallucinations associated with auditory
Kucharska-Pietura, K., David, A., Masiak, K., Phillips, M., 2005. Perception of facial and affective processing deficits? Schizophr. Res. 78 (1), 95–106.
vocal affect by people with schizophrenia in early and late stages of illness. Br. Schirmer, A., Kotz, S.A., 2006. Beyond the right hemisphere: brain mechanisms mediating
J. Psychiatry 187, 523–528. vocal emotional processing. Trends Cogn. Sci. 10 (1), 24–30.
Laurens, K.R., Kiehl, K.A., Ngan, E.T., Liddle, P.F., 2005. Attention orienting dysfunction during Scholten, M.R., Aleman, A., Kahn, R.S., 2008. The processing of emotional prosody and
salient novel stimulus processing in schizophrenia. Schizophr. Res. 75 (2–3), 159–171. semantics in schizophrenia: relationship to gender and IQ. Psychol. Med. 38 (6),
Leitman, D.I., Laukka, P., Juslin, P.N., Saccente, E., Butler, P., Javitt, D.C., 2010. Getting the 887–898.
cue: sensory contributions to auditory emotion recognition impairments in schizo- Shea, T.L., Sergejew, A.A., Burnham, D., Jones, C., Rossell, S.L., Copolov, D.L., Egan, G.F.,
phrenia. Schizophr. Bull. 36 (3), 545–556. 2007. Emotional prosodic processing in auditory hallucinations. Schizophr. Res. 90,
Leitman, D.I., Wolf, D.H., Laukka, P., Ragland, J.D., Valdez, J.N., Turetsky, B.I., Gur, R.E., Gur, 214–220.
R.C., 2011. Not pitch perfect: sensory contributions to affective communication Shenton, M.E., Kikinis, R., Jolesz, F.A., Pollak, S.D., LeMay, M., Wible, C.G., Hokama, H.,
impairment in schizophrenia. Biol. Psychiatry 70 (7), 611–618. Martin, J., Metcalf, D., Coleman, M., et al., 1992. Abnormalities of the left temporal
Lenz, D., Schadow, J., Thaerig, S., Busch, N.A., Herrmann, C.S., 2007. What's that sound? lobe and thought disorder in schizophrenia. A quantitative magnetic resonance
Matches with auditory long-term memory induce gamma activity in human EEG. imaging study. N. Engl. J. Med. 327 (9), 604–612.
Int. J. Psychophysiol. 64 (1), 31–38. Silver, H., Feldman, P., Bilker, W., Gur, R.C., 2003. Working memory deficit as a core neu-
Li, H., Chan, R.C., McAlonan, G.M., Gong, Q.Y., 2010. Facial emotion processing in ropsychological dysfunction in schizophrenia. Am. J. Psychiatry 160 (10), 1809–1816.
schizophrenia: a meta-analysis of functional neuroimaging data. Schizophr. Bull. 36 van Petten, C., 1995. Words and sentences: event-related brain potential measures.
(5), 1029–1039. Psychophysiology 32 (6), 511–525.
Liu, T., Pinheiro, A.P., Deng, G., Nestor, P.G., McCarley, R.W., Niznikiewicz, M.A., 2012. Wechsler, D., 2008. Wechsler Adult Intelligence Scale: Administration and Scoring
Electrophysiological insights into processing nonverbal emotional vocalizations. Manual, 4th edition. The Psychological Corporation, San Antonio.
Neuroreport 23 (2), 108–112. Wheeldon, L., Lahiri, A., 1997. Prosodic units in speech production. J. Mem. Lang. 37 (3),
McCarley, R.W., Faux, S.F., Shenton, M.E., Nestor, P.G., Adams, J., 1991. Event-related po- 356–381.
tentials in schizophrenia: their biological and clinical correlates and a new model of Wilson, M.D., 1998. The MRC psycholinguistic database: machine readable dictionary,
schizophrenic pathophysiology. Schizophr. Res. 4 (2), 209–231. version 2. Behav. Res. Methods Instrum. 20, 6–11.