- Email: [email protected]
P300 and conceptual loosening in normals: An event-related potential correlate of “thought disorder?”
650
BIOL PSYCHIATRY 1992;31:650-660
'\ ,.
P300 and Conceptual Loosening in NormaIs: An Event-Related Potential Correlate of "Thought Disorder?" Philip B. Ward, Stanley V. Catts, and Nathaniel McConaghy
Reduced amplitude of the P300 component of the event-related potential (ERP) has frequently been reported in schizophrenic patients and their first.degree relatives. The present study examined the relationship between this ERP measure of attentional processing and loosening of associations in normal university students (termed "allusive thinking"). Among male subjects, scores reflecting increased conceptual loosening, measured using the Lovibond scoring method for the Goldstein-Scheerer Object Sorting Test (OST), were significantly correlated with smaller P300 emplitude recorded during an auditory target detection task. There **,asno association between OST score and either performance of the target detection task or self-reported psychopathology. It is suggested that reduced P300 amplitude could reflect altered attentional processing in individuals with a constitutional trait f,~ctor of thought disorder.
Introduction One of the most frequently reported electrophysiological findings in schizophrenia is reduced amplitude of a late positive component, termed P300, measured in scalp recordings of event-related potentials (for a review see Pritchard 1986). P300 amplitude has been shown to reflect the allocation of limited capacity attentional resources to task relevant stimuli (Pritchard 1981; Donchin et ai 1986), and is multiplicatively influenced by the subjective probability and task relevance of eliciting events (Donchin and Coins 1988). Some studies examining children of schizophrenic parents (Friedman et ai 1982) and the nonpsychiatrically ill sibliags ~of~.schizophrenics (Saitoh et al 1984) have also found smaller P300 amplitudes in these genetically related populations, although others have failed to find significant P300 amplitude reduction in these populations (Schreiber et al 1991; Friedman et al 1986). Although reduced P300 amplitudes have also been reported in other psychiatric patient samples (e.g., Pfefferbaum et al 1984), the evidence suggests that reduced P300 amplitude may in part reflect a constitutional information processing abnormality in schizophrenia. We have previously reported a preliminary study in male university students in which From the School of Psychiatry, University of New South Wales, Kensington. N.S.W., and Psychiatric Unit, Prince of Wales Hospital, Randwick, N.S.W., Australia. Address reprint requests to Dr. Philip B. Ward, Psychiatric Unit, Prince of Wales Hospital, High Street, Randwick, N.S.W. 2031 Australia. Received June 19, 1990; revised November 8, 1991.
© 1992 Society of Biological Psychiatry
0006-3223/92/$05.00
ERP and Thought Disorder
BIOL PSYCHIATRY 1992;31:650-660
65 I
subjects with high scores oD the Goldstein-Scheerer object sorting test (OST) measure of "conceptual loosening" had significantly reduced P300 amplitudes compared with subjects with low OST scores (Ward et al 1984). McConaghy (1960, 1961, 1989) has proposed that the "loosening of associations" detected by this measure is an index of "allusive thinking," which he defined as a form of conceptual thinking characterized by broader, less focused ideas. McConaghy and Clancy (1968) demonstrated that students with allusive thinking were more likely to have parents with allusive thinking. A similar pattern of familial transmission had also been found in schizophrenics and their parents (McConaghy 1959; Lidz et al 1963; Rosman et al 1964; Carts et al 1986b). As OSTassessed allusive thinking had been shown to be present in a high percentage of schizophrenics and to persist in them following remission of the illness (Lovibond 1954), it was suggested that allusive thinking is a genetically transmitted trait that acted as a predisposition to schizophrenia, though only a minority of individuals with allusive thinking actually develop a schizophrenic illness. It has also been reported that allusive thinking may be associated with high levels of creativity (Tucker et al 1982). Recent studies have demonstrated that impaired attentional processing is a core feature of schizophrenia that shows significant trait characteristics (Nuechterlein and Dawson 1984). Evidence that the attention process of allusive thinking normals was broader and less intense than that of nonallusive thinkers was obtained using verbal tasks (Armstrong and McConaghy 1977; Ward et al 1992b). Allusive thinkers were found to focus less accurately on the meaning of words, leading to more verbose and less specific speech patterns and to more unusual word associations. Such indirect evidence of attentional differences in allusive thinking would be strengthened by ERP measures of attentional processing that are heritable (Polich and Burns 1987) and are less sensitive to cultural factors. The present study was designed to further investigate the association between reduced P300 amplitude and allusive thinking initially reported by Ward and colleagues (1984). In that study, male university students were selected based on a group-administered version of the OST, and a relatively nondemanding task was employed. Subjects silently counted a series of 150 identical auditory tones presented with a regular 2-sec interstimulus interval. As a result, the P300 component was only apparent in averages based on the first 45 stimuli, and was often small in amplitude without a clearly defined peak. As P300 amplitude varies with subjective probability, stimulus salience, and information delivery (Johnson 1986), an auditory oddball task, in which subjects identified an infrequent "target" stimuli, was employed in the present study.
Method
Subjects Students at the University of New South Wales were invited to participate as paid volunteers in a series of cognitive and ERP investigations, which included the administration of the OST and subsequent recording of ERPs in the auditory "oddball" paradigm. OST and ERP data were successfully obtained from 74 subjects, 50 men and 24 women, ranging in age from 18 to 31 (mean 19.55, SD 2.0). As previous work (Carts et al 1986a) suggested gender differences in the relationship between ERPs and allusive thinking, data obtained from male and femak~ subjects were examined separately in this study.
652
toOL
PSYCHIATRY
P.B. Ward et al
1992;31:650--660
Assessment The Symptom Check List 90 (SCL-90), a 90-item self-rated scale measuring the number and intensity of psychiatric symptoms (Derogatis et al 1973), was completed by all subjects. The OST was administered in the standardized form (Lovibond 1966). Briefly, subjects were presented with familiar commonplace items (e.g., cutlery, tools, toys) and asked why various subgroups "belong" together. In the first part of the test, the subject was presented with a single object (e.g., a dinner fork) and asked firstly to indicate all those objects that "belong" with it and secondly the reason they belong together. Seven individual objects were presented. In the second part, the subject was shown a group of ,,r,,,.,o ~ , . .m .,,~,,,I ,t,;,~,.t~ ~.a ,,,~ ~ r , ~ked why the objects belong together. Twelve groups were presented. Subjects' entire verbal response to each item was taped and later transcribed for scoring by an independent rater. Scores were determined for the degree to which the response exhibited syncretistic, fabulated, arbitrary, inessential, chained, or symbolic thinking. Ez~.iaitem was scored on a 4-point scale for the "loosest" response offered by the subject. The transcripts were scored by an experienced rater (NMc ~), who assisted in the development of the Lovibond scoring system and has scored OST protocols for over 25 years. He remained blind to ERP and other cognitive test results and scored all protocols in a single group at the completion of the study.
Auditory ERP Paradigm The ERP paradigm used was a standard auditory "oddball" design, in which subjects listened to a sequence of 300 tones, presented at an intensity of 85 dB SPL through a loudspeaker situated overhead, and made a button press response to infrequent target stimuli (n = 36). All tones were of 100 msec duration with rise and fall times of 10 msec, presented every 2 sec. The frequent (nontarget) stimulus was an 800-Hz tone, and the infrequent (Target) stimulus was a 1200-Hz tone. These were delivered in a pseudorandom sequence such that there was a probability of 0. ! that any tone was a Target. The sequence used was the same f~r all subjects, who were instructed to make a button press with the preferred hand to the 1200-Hz Targets. A short practice run preceded the recording to familiarize them with the auditory discrimination task. Subjects were seated in a semirecumbent position, in a dimly lit, sound-attenuated laboratory for the recording. Silver/silver chloride electrodes were attached with collodion to the scalp at Fz, Cz, Pz, according to the International 10120 system, after cleaning and gentle abrasion to reduce electrode impedance below 5 kfl. The electrooculogram (EOG) was recorded from an electrode located 2 cm below the left eye, and all electrodes were referenced to linked mastoids, with a forearm ground. Both the EEG and EOG were amplified with a bandpa.~s of 0.2-35 Hz, and digitized at 250 Hz for 1.024 sec following each stimulus. Subjects were asked to fixate on a red disk during the 10-rain run, and to avoid unnecessary blinking. Averaging was carried out off-line, with automatic rejection of epochs contaminated by eye movements, to create separate averages for target and background stimuli for the three EEG derivations and the EnG. Reaction times were recorded via a Schmitt trigger. ERP measures were scored using an automatic peak detection program, which identified the most positive peak between 235 and 450 ms as P300 and measured the amplitude
ERP and Thought Disorder
BIOLPSYCHIATRY 1992;31:650-660
653
Table !. Spearman Correlationsbetween OST Score and P300 Amplitude
Electrode location Frontal (Fz) Central (Cz) Parietal (Pz)
Male subjects (n = 50)
Female subjects (n ffi 24)
P300 amplitude Nontarget Target
P300 amplitude Nontarget Target
0.05 0.11 0.12
- 0.11 - 0.22 - 0.25 a
0.14 0.28 0.25
- 0.18 0.07 0.17
op < 0.05.
and latency of this peak relative to the first data point for Target and Nont~rget averages from each electrode site. Averages for frequent stimuli were based on approximately 264 events, while approximately 36 events were used to construct averages for infrequent stimuli.
Results The distribution of OST scores was significantly skewed, so all tests of significance involving OST score were made using nonparametric statistics (i.e., Spearman rank order correlations and KruskaI-WaUis analysis of variance).
OST and SCL-90 Scores Male and female subjects did not differ significantly in OST scores (males: mean OST score = 8.04, SD = 6.03; females: mean OST score = 5.25, SD = 5 25; ~2 ,_ 3.35, NS). Correlations were calculated between OST score and the three summary measures provided by the SCL-90: Global Severity Index (GSI), Positive Symptom Distress Index (PSDI), and Positive Symptom Total (PST). The correlations did not approach significance in either male or female groups. Behavioral Data The behavioral data indicate that all subjects performed the ERP task with a high degree of accuracy. The mean percel~tage of correctly identified Targets ("Hits") was 99.7%, and no subject made more than three button presses to the Nontargets ("False Alarms"). Mean reaction time was 425.9 ms (SD 100.2), and the variability of reaction time within subjects was examined by calculating the within-subject standard deviation. OST score was not significantly correlated with any behavioral performance measure (% hits, false alarms, reaction time, within-subject reaction time variability) for either male or female subjects. OST Score and P300 Amplitude Correlations between OST score and P300 amplitude measured at the three scalp derivations in both Target and Nontarget ERPs were computed separately for male and female subjects (Table 1). In male subjects, OST score was negatively correlated with P300 amplitude to Targets at Pz (rS = -0.25, p = 0.04; Figure 1). No consistent relationship
654
B[OL PSYCHIATRY
P.B. Ward et al
1992,31:650-660
40
•
•
3O
o
up
• •
•
~J •
•
i
|
•
•
•
,
•
0
•
•
•
:
•
Figure 1. Scattergram of relationship between Target !'300 amplitude at Pz and OST scores in m~]e ~bjects (n = 50).
•
•
! '.~,
,/
e
J
r
I
I
1
5
10
15
20
25
OST Score
40
ii B
3o
|
>
eL
.
Figure 2. Scattergram of relationship between Target P300 amplitude at Pz and OST score in female subjects (n = 24).
~'"
2, ¸
*l
1
l
I
I
I0
15
20
25
OST Score
ERP and Thought Disorder
BIOLPSYCHIATRY 1992;31:650-660
655
between P300 amplitude to Targets and OST was evident in female subjects (Figure 2). A negative correlation was also found between OST score and P300 latency for Nontarget Pz data in male subjects ( r S = -0.28, p = 0.02). To illustrate the significant association between P300 amplitude and OST scores in men, grand average ERPs for the 11 highest scoring (OST > 10) and lowest scoring (OST < 4) male subjects were formed, and are presented in Figure 3. The marked difference in P300 amplitude to Targets and Nontargets in the male low OST scorers grand average ERP is much less apparent in the high OST scorers grand average ERP. This effect is more readily seen in the Pz site waveforms than at the Fz and Cz sites. To evaluate this observation, Nontarget P300 amplitude was ~;ubtraeted from Target P300 amplitude for each subject at each electrode site, and correlations were calculated between the resulting P300 amplitude difference score and OST score (Table 2). Significant negative correlations with OST were obtained in male, but not female subjects, at Pz (p - 0.009) and Cz (p - 0.025) electrodes (Figures 4 and 5). Discussion This study provides further evidence that increased levels of thought disorder in normals, when measured using the OST, are not associated with marked psychopathology, as assessed using the SCL-90. This finding supports McConaghy's (1960, 1961) hypothesis that allusive thinking is a normal mode of conceptual thinking that, while it may reflect a predisposition to schizophrenia, does not indicate the presence of an undiagnosed psychiatric illness in normals. Previous evidence of reduced P300 amplitudes in high OST scorers was found in an independent group of male university students (Ward et al 1984). The present study replicates this association, at least in male subjects. The absence of a significant relationship between P300 amplitude and OST scores in women could be a chance negative finding or type II error, resulting from the smaller number of female subjects tested and the trend towards lower OST scores among women in this study. Alternatively, it may reflect a true sex difference, as has been suggested by previous ERP studies of both allusive thinking (Carts et al 1986a) and other cognitive variables (Molfese et al 1979). The possibility that the interaction between attention, as indexed by i'300 amplitude, and conceptual loosening, as indexed by OST score differs in men and women should be addressed in future studies. In this study, the auditory discrimination required to carry out the task was relatively easy, and was performed accurately by all subjects. There was no evidence that subjects with high OST scores found the task more difficult; their reaction times were as accurate, fast, and consistent as those recorded in low OST scorers. Despite this, the ERP data suggest differences in the attentional processes employed by male allusive thinkers in this task. High OST scores were associated with reduced differentiation between the taskrelevant Target tones and the less salient Nontarget stimuli, as shown by reduced P300 amplitude of the Target ERPs and smaller differences in P300 amplitudes between Target and Nontarget ERPS. P300 amplitude reduction has been reported when normal subjects were instructed to passively ignore the stimuli, or have actively undertaken distracting tasks (Polich 1986). Holcomb and colleagues (1985) also found a smaller difference between I'300 amplitude to Target and Nontarget stimuli in boys with attention deficit disorder. These results
656
BIOLPSYCHIATRY
P.B. Ward et al
1992;3 ! :650-660
NON-TARGET ......
TARGET LOV OST MALES ,
HIGH OST MALES
,--"-------
EoG ,, . ~ ~ . . . . . . . .
_...:.._.
Fz
Fz
VV
Vv k y" Cz
,'
I "L' I%,'1"
".," . . . . .
Cz
+ 10uV
-10uV
-%OuV
500 I
I
I
1000 |
|
m~11~secsnQs
|
i
530 |
,
I
I
I
i
1000 i
,5
|
i
i
~n;l:isecsncs
Figure 3. Grand average ERPs for male subjects with Low OST scores (OST < 4; n -- 1 I) and High OST scores (OST > 10; n -- ll).
ERP and Thought Disorder
BIOL
PSYCHIATRY
657
1992;31:650--660
Table 2. Spearman Correlations between OST Score and P300 Amplitude Difference Scores (Target-Nontarget) Electrode location
Male subjects (n = 50)
Female subjects (n = 24)
- 0.05
- O.30
- 0.29"
- 0.03
Frontal (Fz) Central (Cz) Parietal
-
0.38b
-
0.04
(Pz)
,p < 0.05. bp <
O.Ol.
support the hypothesis that allusive thinking reflects a less focused attentional process with reduced inhibition of irrelevant stimuli. The present data constitute a replication and extension of the finding of altered information processing, as indexed by difference in P300 amplitude, associated with allusive thinking. The similarity between the ERP findings in allusive thinkers and those reported in schizophrenic patients suggest that further research is warranted to elucidate the relationship between allusive thinking in healthy individuals and thought-disorder in schizophrenic patients. ERP paradigms that place greater attentional demands on the subject (such as more complex dichotic listening tasks) are currently being employed to test schizophrenic patients (Michie et el 1990; Ward et al 1991a) and healthy controls (McConaghy et al, submitted) to further elucidate the relationship among P300 amplitude, thought disorder, and allusive thinking.
25
20
|
> ::k
•
v
•
0
•
•
~A
(.3 Z L~ ,v
•
,
o°
15 Figure 4. Scattergram of relationship between P300 amplitude difference (Target-Nontarget) at Pz and OST scores in male subjects (n --- 50).
L~
10 < C C
@ 0 •
•
i
I
T
I
0
5
10
15
OST Score
I
20
I
25
658
BIOLPSYCHIATRY
P.B. Ward et al
1992;3 ! :650--660
30
25
@
@ @
::k
2O (3 z
@
@
@ _
a
@
,
Figure 5. Scattergram of relationship between P300 a~:plitude difference (Target-Nontarget) at Pz and OST scores in female subjects (n = 24).
@ @
15
< 0 0
1o
@
I @ @
g.
0
0
5
10
I
I
15
2O
__
__J
25
OST Score
This researchwas supportedby the National Health and MedicalResearchCouncilof Australia, the Rebecca L. Cooper Medical ResearchFoundation, and by a ResearchFellowship to SVC from the NSW Instituteof Psychiatry.The assistanceof MichaelAmlstrong, AllisonFox, John Garvey, and Ken Norcross is gratefully acknowledged.
References Armstrong MS, McConaghy N (1977): Allusive thinking, the word halo, and verbosity. Psychol Med 7:439-445. Catts SV, Armstrong MS, Ward PB, McConaghy N (1986a): Reduced P200 latency and allusive thinking: An auditory evoked potential index of a cognitive predisposition to schizophrenia. I~r J Neurosci 30:173-179. Carts SV, Ward PB, Armstrong MS, McConaghy N (1986b): Auditory evoked potentials and allusive thinking in the parents of schizophrenics: A preliminary report. Electroencephalogr Clin Neurophysiol 38(suppl): 443-445. I)erogatis LR, Lipman RS, Covi L (19"/3): The SCL-90: An outpatient psychiatric rating scale. Psychopharraacol Bull 9:13-28. !)onchin E, Coles MGH (I 988): Is the P300 c~mponem a manifestation of context updating.'?Behav Brain Sci 11:357-374. Donchin E, Karis D, Bashore TR, Coles MGH, Gratton G (1986): Cognitive psychophysiology and human information processing. In Coles MGH, Donchin E, Porges SW (eds), P~ychophysiology: Systems, Processes, and Applications. New York: Guilford, pp 244-267. Friedman D, Vaughan HG Jr, Erlenmeyer-Kimling L (1982): Cognitive brain potentials in children at risk for schizophrenia: Preliminary findings. Schizophr Bull 8:514-531. Friedman D, Cornblatt B, Vaughan HG Jr, Erlenmeyer-Kimling L (1986): Event-related pote~tials in children at risk for schizophrenia during two versions of the continuous performance test. Psychiatry Res 18:161-177.
ERP and Thoug,~t Disorder
BIOLPSYCHIATRY 1992;31:650-660
659
Holcomb PJ, Ackerman lrl ', Dykman RA ( 1985): Cognitive event-related brain potentials in children with attention and reading deficits. Psychophysiology 22:656-667. Johnson R (1986): A triarchic model of P300 amplitude. Psychophysiology 23:367-384. Lidz T, Wild C, Schafer S, Rosman B, Fleck S 0963): Thought disorders in the parents of schizophrenic patients: A study utilizing the Object Sorting Test. J Psychiatr Res 1:193-200. Lovibond SH (1954): The Object Sorting Test and conceptual thinking in schizophrenia. Aust J Psychol 6:52-70. Lovibond SH (1966): Interim Manual for the Object Sorting Test. Melbourne: Australian Council for Educational Research. McConaghy N (1959): The use of an object scoring test in elucidating the hereditary factor in schizophrenia. J Neurol Neurosurg Psychiatry 22:243-246. McConaghy N (1960): Modes of abstract thinking and psychosis. Am J Psychiatry 117:106-110 McConaghy N (1961): The measurement of an inhibitory process in human higher nervous activity: Its relation to allusive thinking and fatigue. Am J Psychiatry 118:i25-132. McConaghy N (1989): Thought disorder or allusive thinking in the relative of schizophrenics? A response to Callahan, Madsen, Saccuzzo, and Romney. J Nerv blent Dis 177:729-734. McConaghy N, Clancy M (1968): Familial relationships of allusive thinking in university students and their parents. Br J Psychiatry 114:1079-1087. McConaghy N, Carts SV, Michie PT, Fox A, Ward PB (submitted): P300 indexes a cognitive vulnerability to schizophrenia in normal subjects. Michie PT, Fox AM, Ward PB, Catts SV, McConaghy N (1990): ERP indices of selective attention and cortical lateralization in schizophrenia. Psychophysiology 27:209-227. Molfese DL, Papanicolaou A, Hess TM, Molfese VJ (1979): Neuroolectrical correlates of semantic processes, in Begleiter H (ed), Evoked Brain Potentials and Behavior. New York: Plenum, pp 89-106. Nuechterlein KH, Dawson ME (1984): Information processing and attentional functioning in the developmental course of schizophrenk disorders. Schizophr Bull 10:160-203. Pfefferbaum A, Wenegrat BG, Ford JM, Roth WT, Kopell BS (1984): Clinical application of the P3 component of event-related potentials. II Dementia, depression and schizophrenia. Electroencephalogr Ciin Neurophysiol 59:104-124. Polich J (1986): Attention, probability, and task demands as determinants of P300 latency from auditory stimuli. Electroencephalogr Clin Neurophysio! 63:25 !-259. Polich J, Burns T (1987): P300 from identical twins. Neuropsychologia 25:299-304. Pritchard WS (1981): Psychophysiology of P300. Psychol Bull 89:506-540. P:i~chard WS (1986): Cognitive event-related potential correlates of schizophrenia. Psychol Bull 100:43-66. Rapaport D (1945): Diagnostic Psychological Testing. Chicago: Year Book. Rosman B, Wild C, Ricci J, Fleck S, Lidz T (1964): Thought disorders in the parents of schizophrenic patients: A further study utilizing the Object Sorting Test. J Psychiatr Res 2:2~ !-221. Saitoh O, Niwa S, Hiramatsu K, Kameyama T, Rymar K, Itoh K (1984): Abnormalities in late positive components of event-related potentials may reflect a genetic predisposition to schizophrenia. Biol Psychiatry 19:293-303. Schreiber H, Stolz-Born G, Rothmeier J, Kornhuber A, Kornhuber HH, Born J (1991): Endogenous event-related brain potentials and psychometric performance in children at risk for schizophrenia. Biol Psychiatry 30:177-189. Tucker PK, Rothwell SJ, Armstrong MS, McConaghy N (1982): Creativity, divergent and allusive thinking in students and visual artists. Psychol Med 12:835-841.
BIOLPSYCHIATRY
P.B. Ward et al
1992;31:650-660
Ward PB, CaRs SV, Armstrong MA, McConaghy N (1984): P300 and psychiatric vulnerability in university students. Ann NY Acad Sci 425:645--652. Ward PB, Catts SV, Fox AM, Michie PT, McConaghy N (1991a): Audi:o~, selective attention and event-related potentials in schizophrenia. Br J Psychiatry 158:534--539. Ward PB, McConaghy N, CaRs SV (1991b): Word association and measures of psychosis proneness in university students. Person Indiv Diff 12:473-480.
BIOL PSYCHIATRY 1992;31:650-660
'\ ,.
P300 and Conceptual Loosening in NormaIs: An Event-Related Potential Correlate of "Thought Disorder?" Philip B. Ward, Stanley V. Catts, and Nathaniel McConaghy
Reduced amplitude of the P300 component of the event-related potential (ERP) has frequently been reported in schizophrenic patients and their first.degree relatives. The present study examined the relationship between this ERP measure of attentional processing and loosening of associations in normal university students (termed "allusive thinking"). Among male subjects, scores reflecting increased conceptual loosening, measured using the Lovibond scoring method for the Goldstein-Scheerer Object Sorting Test (OST), were significantly correlated with smaller P300 emplitude recorded during an auditory target detection task. There **,asno association between OST score and either performance of the target detection task or self-reported psychopathology. It is suggested that reduced P300 amplitude could reflect altered attentional processing in individuals with a constitutional trait f,~ctor of thought disorder.
Introduction One of the most frequently reported electrophysiological findings in schizophrenia is reduced amplitude of a late positive component, termed P300, measured in scalp recordings of event-related potentials (for a review see Pritchard 1986). P300 amplitude has been shown to reflect the allocation of limited capacity attentional resources to task relevant stimuli (Pritchard 1981; Donchin et ai 1986), and is multiplicatively influenced by the subjective probability and task relevance of eliciting events (Donchin and Coins 1988). Some studies examining children of schizophrenic parents (Friedman et ai 1982) and the nonpsychiatrically ill sibliags ~of~.schizophrenics (Saitoh et al 1984) have also found smaller P300 amplitudes in these genetically related populations, although others have failed to find significant P300 amplitude reduction in these populations (Schreiber et al 1991; Friedman et al 1986). Although reduced P300 amplitudes have also been reported in other psychiatric patient samples (e.g., Pfefferbaum et al 1984), the evidence suggests that reduced P300 amplitude may in part reflect a constitutional information processing abnormality in schizophrenia. We have previously reported a preliminary study in male university students in which From the School of Psychiatry, University of New South Wales, Kensington. N.S.W., and Psychiatric Unit, Prince of Wales Hospital, Randwick, N.S.W., Australia. Address reprint requests to Dr. Philip B. Ward, Psychiatric Unit, Prince of Wales Hospital, High Street, Randwick, N.S.W. 2031 Australia. Received June 19, 1990; revised November 8, 1991.
© 1992 Society of Biological Psychiatry
0006-3223/92/$05.00
ERP and Thought Disorder
BIOL PSYCHIATRY 1992;31:650-660
65 I
subjects with high scores oD the Goldstein-Scheerer object sorting test (OST) measure of "conceptual loosening" had significantly reduced P300 amplitudes compared with subjects with low OST scores (Ward et al 1984). McConaghy (1960, 1961, 1989) has proposed that the "loosening of associations" detected by this measure is an index of "allusive thinking," which he defined as a form of conceptual thinking characterized by broader, less focused ideas. McConaghy and Clancy (1968) demonstrated that students with allusive thinking were more likely to have parents with allusive thinking. A similar pattern of familial transmission had also been found in schizophrenics and their parents (McConaghy 1959; Lidz et al 1963; Rosman et al 1964; Carts et al 1986b). As OSTassessed allusive thinking had been shown to be present in a high percentage of schizophrenics and to persist in them following remission of the illness (Lovibond 1954), it was suggested that allusive thinking is a genetically transmitted trait that acted as a predisposition to schizophrenia, though only a minority of individuals with allusive thinking actually develop a schizophrenic illness. It has also been reported that allusive thinking may be associated with high levels of creativity (Tucker et al 1982). Recent studies have demonstrated that impaired attentional processing is a core feature of schizophrenia that shows significant trait characteristics (Nuechterlein and Dawson 1984). Evidence that the attention process of allusive thinking normals was broader and less intense than that of nonallusive thinkers was obtained using verbal tasks (Armstrong and McConaghy 1977; Ward et al 1992b). Allusive thinkers were found to focus less accurately on the meaning of words, leading to more verbose and less specific speech patterns and to more unusual word associations. Such indirect evidence of attentional differences in allusive thinking would be strengthened by ERP measures of attentional processing that are heritable (Polich and Burns 1987) and are less sensitive to cultural factors. The present study was designed to further investigate the association between reduced P300 amplitude and allusive thinking initially reported by Ward and colleagues (1984). In that study, male university students were selected based on a group-administered version of the OST, and a relatively nondemanding task was employed. Subjects silently counted a series of 150 identical auditory tones presented with a regular 2-sec interstimulus interval. As a result, the P300 component was only apparent in averages based on the first 45 stimuli, and was often small in amplitude without a clearly defined peak. As P300 amplitude varies with subjective probability, stimulus salience, and information delivery (Johnson 1986), an auditory oddball task, in which subjects identified an infrequent "target" stimuli, was employed in the present study.
Method
Subjects Students at the University of New South Wales were invited to participate as paid volunteers in a series of cognitive and ERP investigations, which included the administration of the OST and subsequent recording of ERPs in the auditory "oddball" paradigm. OST and ERP data were successfully obtained from 74 subjects, 50 men and 24 women, ranging in age from 18 to 31 (mean 19.55, SD 2.0). As previous work (Carts et al 1986a) suggested gender differences in the relationship between ERPs and allusive thinking, data obtained from male and femak~ subjects were examined separately in this study.
652
toOL
PSYCHIATRY
P.B. Ward et al
1992;31:650--660
Assessment The Symptom Check List 90 (SCL-90), a 90-item self-rated scale measuring the number and intensity of psychiatric symptoms (Derogatis et al 1973), was completed by all subjects. The OST was administered in the standardized form (Lovibond 1966). Briefly, subjects were presented with familiar commonplace items (e.g., cutlery, tools, toys) and asked why various subgroups "belong" together. In the first part of the test, the subject was presented with a single object (e.g., a dinner fork) and asked firstly to indicate all those objects that "belong" with it and secondly the reason they belong together. Seven individual objects were presented. In the second part, the subject was shown a group of ,,r,,,.,o ~ , . .m .,,~,,,I ,t,;,~,.t~ ~.a ,,,~ ~ r , ~ked why the objects belong together. Twelve groups were presented. Subjects' entire verbal response to each item was taped and later transcribed for scoring by an independent rater. Scores were determined for the degree to which the response exhibited syncretistic, fabulated, arbitrary, inessential, chained, or symbolic thinking. Ez~.iaitem was scored on a 4-point scale for the "loosest" response offered by the subject. The transcripts were scored by an experienced rater (NMc ~), who assisted in the development of the Lovibond scoring system and has scored OST protocols for over 25 years. He remained blind to ERP and other cognitive test results and scored all protocols in a single group at the completion of the study.
Auditory ERP Paradigm The ERP paradigm used was a standard auditory "oddball" design, in which subjects listened to a sequence of 300 tones, presented at an intensity of 85 dB SPL through a loudspeaker situated overhead, and made a button press response to infrequent target stimuli (n = 36). All tones were of 100 msec duration with rise and fall times of 10 msec, presented every 2 sec. The frequent (nontarget) stimulus was an 800-Hz tone, and the infrequent (Target) stimulus was a 1200-Hz tone. These were delivered in a pseudorandom sequence such that there was a probability of 0. ! that any tone was a Target. The sequence used was the same f~r all subjects, who were instructed to make a button press with the preferred hand to the 1200-Hz Targets. A short practice run preceded the recording to familiarize them with the auditory discrimination task. Subjects were seated in a semirecumbent position, in a dimly lit, sound-attenuated laboratory for the recording. Silver/silver chloride electrodes were attached with collodion to the scalp at Fz, Cz, Pz, according to the International 10120 system, after cleaning and gentle abrasion to reduce electrode impedance below 5 kfl. The electrooculogram (EOG) was recorded from an electrode located 2 cm below the left eye, and all electrodes were referenced to linked mastoids, with a forearm ground. Both the EEG and EOG were amplified with a bandpa.~s of 0.2-35 Hz, and digitized at 250 Hz for 1.024 sec following each stimulus. Subjects were asked to fixate on a red disk during the 10-rain run, and to avoid unnecessary blinking. Averaging was carried out off-line, with automatic rejection of epochs contaminated by eye movements, to create separate averages for target and background stimuli for the three EEG derivations and the EnG. Reaction times were recorded via a Schmitt trigger. ERP measures were scored using an automatic peak detection program, which identified the most positive peak between 235 and 450 ms as P300 and measured the amplitude
ERP and Thought Disorder
BIOLPSYCHIATRY 1992;31:650-660
653
Table !. Spearman Correlationsbetween OST Score and P300 Amplitude
Electrode location Frontal (Fz) Central (Cz) Parietal (Pz)
Male subjects (n = 50)
Female subjects (n ffi 24)
P300 amplitude Nontarget Target
P300 amplitude Nontarget Target
0.05 0.11 0.12
- 0.11 - 0.22 - 0.25 a
0.14 0.28 0.25
- 0.18 0.07 0.17
op < 0.05.
and latency of this peak relative to the first data point for Target and Nont~rget averages from each electrode site. Averages for frequent stimuli were based on approximately 264 events, while approximately 36 events were used to construct averages for infrequent stimuli.
Results The distribution of OST scores was significantly skewed, so all tests of significance involving OST score were made using nonparametric statistics (i.e., Spearman rank order correlations and KruskaI-WaUis analysis of variance).
OST and SCL-90 Scores Male and female subjects did not differ significantly in OST scores (males: mean OST score = 8.04, SD = 6.03; females: mean OST score = 5.25, SD = 5 25; ~2 ,_ 3.35, NS). Correlations were calculated between OST score and the three summary measures provided by the SCL-90: Global Severity Index (GSI), Positive Symptom Distress Index (PSDI), and Positive Symptom Total (PST). The correlations did not approach significance in either male or female groups. Behavioral Data The behavioral data indicate that all subjects performed the ERP task with a high degree of accuracy. The mean percel~tage of correctly identified Targets ("Hits") was 99.7%, and no subject made more than three button presses to the Nontargets ("False Alarms"). Mean reaction time was 425.9 ms (SD 100.2), and the variability of reaction time within subjects was examined by calculating the within-subject standard deviation. OST score was not significantly correlated with any behavioral performance measure (% hits, false alarms, reaction time, within-subject reaction time variability) for either male or female subjects. OST Score and P300 Amplitude Correlations between OST score and P300 amplitude measured at the three scalp derivations in both Target and Nontarget ERPs were computed separately for male and female subjects (Table 1). In male subjects, OST score was negatively correlated with P300 amplitude to Targets at Pz (rS = -0.25, p = 0.04; Figure 1). No consistent relationship
654
B[OL PSYCHIATRY
P.B. Ward et al
1992,31:650-660
40
•
•
3O
o
up
• •
•
~J •
•
i
|
•
•
•
,
•
0
•
•
•
:
•
Figure 1. Scattergram of relationship between Target !'300 amplitude at Pz and OST scores in m~]e ~bjects (n = 50).
•
•
! '.~,
,/
e
J
r
I
I
1
5
10
15
20
25
OST Score
40
ii B
3o
|
>
eL
.
Figure 2. Scattergram of relationship between Target P300 amplitude at Pz and OST score in female subjects (n = 24).
~'"
2, ¸
*l
1
l
I
I
I0
15
20
25
OST Score
ERP and Thought Disorder
BIOLPSYCHIATRY 1992;31:650-660
655
between P300 amplitude to Targets and OST was evident in female subjects (Figure 2). A negative correlation was also found between OST score and P300 latency for Nontarget Pz data in male subjects ( r S = -0.28, p = 0.02). To illustrate the significant association between P300 amplitude and OST scores in men, grand average ERPs for the 11 highest scoring (OST > 10) and lowest scoring (OST < 4) male subjects were formed, and are presented in Figure 3. The marked difference in P300 amplitude to Targets and Nontargets in the male low OST scorers grand average ERP is much less apparent in the high OST scorers grand average ERP. This effect is more readily seen in the Pz site waveforms than at the Fz and Cz sites. To evaluate this observation, Nontarget P300 amplitude was ~;ubtraeted from Target P300 amplitude for each subject at each electrode site, and correlations were calculated between the resulting P300 amplitude difference score and OST score (Table 2). Significant negative correlations with OST were obtained in male, but not female subjects, at Pz (p - 0.009) and Cz (p - 0.025) electrodes (Figures 4 and 5). Discussion This study provides further evidence that increased levels of thought disorder in normals, when measured using the OST, are not associated with marked psychopathology, as assessed using the SCL-90. This finding supports McConaghy's (1960, 1961) hypothesis that allusive thinking is a normal mode of conceptual thinking that, while it may reflect a predisposition to schizophrenia, does not indicate the presence of an undiagnosed psychiatric illness in normals. Previous evidence of reduced P300 amplitudes in high OST scorers was found in an independent group of male university students (Ward et al 1984). The present study replicates this association, at least in male subjects. The absence of a significant relationship between P300 amplitude and OST scores in women could be a chance negative finding or type II error, resulting from the smaller number of female subjects tested and the trend towards lower OST scores among women in this study. Alternatively, it may reflect a true sex difference, as has been suggested by previous ERP studies of both allusive thinking (Carts et al 1986a) and other cognitive variables (Molfese et al 1979). The possibility that the interaction between attention, as indexed by i'300 amplitude, and conceptual loosening, as indexed by OST score differs in men and women should be addressed in future studies. In this study, the auditory discrimination required to carry out the task was relatively easy, and was performed accurately by all subjects. There was no evidence that subjects with high OST scores found the task more difficult; their reaction times were as accurate, fast, and consistent as those recorded in low OST scorers. Despite this, the ERP data suggest differences in the attentional processes employed by male allusive thinkers in this task. High OST scores were associated with reduced differentiation between the taskrelevant Target tones and the less salient Nontarget stimuli, as shown by reduced P300 amplitude of the Target ERPs and smaller differences in P300 amplitudes between Target and Nontarget ERPS. P300 amplitude reduction has been reported when normal subjects were instructed to passively ignore the stimuli, or have actively undertaken distracting tasks (Polich 1986). Holcomb and colleagues (1985) also found a smaller difference between I'300 amplitude to Target and Nontarget stimuli in boys with attention deficit disorder. These results
656
BIOLPSYCHIATRY
P.B. Ward et al
1992;3 ! :650-660
NON-TARGET ......
TARGET LOV OST MALES ,
HIGH OST MALES
,--"-------
EoG ,, . ~ ~ . . . . . . . .
_...:.._.
Fz
Fz
VV
Vv k y" Cz
,'
I "L' I%,'1"
".," . . . . .
Cz
+ 10uV
-10uV
-%OuV
500 I
I
I
1000 |
|
m~11~secsnQs
|
i
530 |
,
I
I
I
i
1000 i
,5
|
i
i
~n;l:isecsncs
Figure 3. Grand average ERPs for male subjects with Low OST scores (OST < 4; n -- 1 I) and High OST scores (OST > 10; n -- ll).
ERP and Thought Disorder
BIOL
PSYCHIATRY
657
1992;31:650--660
Table 2. Spearman Correlations between OST Score and P300 Amplitude Difference Scores (Target-Nontarget) Electrode location
Male subjects (n = 50)
Female subjects (n = 24)
- 0.05
- O.30
- 0.29"
- 0.03
Frontal (Fz) Central (Cz) Parietal
-
0.38b
-
0.04
(Pz)
,p < 0.05. bp <
O.Ol.
support the hypothesis that allusive thinking reflects a less focused attentional process with reduced inhibition of irrelevant stimuli. The present data constitute a replication and extension of the finding of altered information processing, as indexed by difference in P300 amplitude, associated with allusive thinking. The similarity between the ERP findings in allusive thinkers and those reported in schizophrenic patients suggest that further research is warranted to elucidate the relationship between allusive thinking in healthy individuals and thought-disorder in schizophrenic patients. ERP paradigms that place greater attentional demands on the subject (such as more complex dichotic listening tasks) are currently being employed to test schizophrenic patients (Michie et el 1990; Ward et al 1991a) and healthy controls (McConaghy et al, submitted) to further elucidate the relationship among P300 amplitude, thought disorder, and allusive thinking.
25
20
|
> ::k
•
v
•
0
•
•
~A
(.3 Z L~ ,v
•
,
o°
15 Figure 4. Scattergram of relationship between P300 amplitude difference (Target-Nontarget) at Pz and OST scores in male subjects (n --- 50).
L~
10 < C C
@ 0 •
•
i
I
T
I
0
5
10
15
OST Score
I
20
I
25
658
BIOLPSYCHIATRY
P.B. Ward et al
1992;3 ! :650--660
30
25
@
@ @
::k
2O (3 z
@
@
@ _
a
@
,
Figure 5. Scattergram of relationship between P300 a~:plitude difference (Target-Nontarget) at Pz and OST scores in female subjects (n = 24).
@ @
15
< 0 0
1o
@
I @ @
g.
0
0
5
10
I
I
15
2O
__
__J
25
OST Score
This researchwas supportedby the National Health and MedicalResearchCouncilof Australia, the Rebecca L. Cooper Medical ResearchFoundation, and by a ResearchFellowship to SVC from the NSW Instituteof Psychiatry.The assistanceof MichaelAmlstrong, AllisonFox, John Garvey, and Ken Norcross is gratefully acknowledged.
References Armstrong MS, McConaghy N (1977): Allusive thinking, the word halo, and verbosity. Psychol Med 7:439-445. Catts SV, Armstrong MS, Ward PB, McConaghy N (1986a): Reduced P200 latency and allusive thinking: An auditory evoked potential index of a cognitive predisposition to schizophrenia. I~r J Neurosci 30:173-179. Carts SV, Ward PB, Armstrong MS, McConaghy N (1986b): Auditory evoked potentials and allusive thinking in the parents of schizophrenics: A preliminary report. Electroencephalogr Clin Neurophysiol 38(suppl): 443-445. I)erogatis LR, Lipman RS, Covi L (19"/3): The SCL-90: An outpatient psychiatric rating scale. Psychopharraacol Bull 9:13-28. !)onchin E, Coles MGH (I 988): Is the P300 c~mponem a manifestation of context updating.'?Behav Brain Sci 11:357-374. Donchin E, Karis D, Bashore TR, Coles MGH, Gratton G (1986): Cognitive psychophysiology and human information processing. In Coles MGH, Donchin E, Porges SW (eds), P~ychophysiology: Systems, Processes, and Applications. New York: Guilford, pp 244-267. Friedman D, Vaughan HG Jr, Erlenmeyer-Kimling L (1982): Cognitive brain potentials in children at risk for schizophrenia: Preliminary findings. Schizophr Bull 8:514-531. Friedman D, Cornblatt B, Vaughan HG Jr, Erlenmeyer-Kimling L (1986): Event-related pote~tials in children at risk for schizophrenia during two versions of the continuous performance test. Psychiatry Res 18:161-177.
ERP and Thoug,~t Disorder
BIOLPSYCHIATRY 1992;31:650-660
659
Holcomb PJ, Ackerman lrl ', Dykman RA ( 1985): Cognitive event-related brain potentials in children with attention and reading deficits. Psychophysiology 22:656-667. Johnson R (1986): A triarchic model of P300 amplitude. Psychophysiology 23:367-384. Lidz T, Wild C, Schafer S, Rosman B, Fleck S 0963): Thought disorders in the parents of schizophrenic patients: A study utilizing the Object Sorting Test. J Psychiatr Res 1:193-200. Lovibond SH (1954): The Object Sorting Test and conceptual thinking in schizophrenia. Aust J Psychol 6:52-70. Lovibond SH (1966): Interim Manual for the Object Sorting Test. Melbourne: Australian Council for Educational Research. McConaghy N (1959): The use of an object scoring test in elucidating the hereditary factor in schizophrenia. J Neurol Neurosurg Psychiatry 22:243-246. McConaghy N (1960): Modes of abstract thinking and psychosis. Am J Psychiatry 117:106-110 McConaghy N (1961): The measurement of an inhibitory process in human higher nervous activity: Its relation to allusive thinking and fatigue. Am J Psychiatry 118:i25-132. McConaghy N (1989): Thought disorder or allusive thinking in the relative of schizophrenics? A response to Callahan, Madsen, Saccuzzo, and Romney. J Nerv blent Dis 177:729-734. McConaghy N, Clancy M (1968): Familial relationships of allusive thinking in university students and their parents. Br J Psychiatry 114:1079-1087. McConaghy N, Carts SV, Michie PT, Fox A, Ward PB (submitted): P300 indexes a cognitive vulnerability to schizophrenia in normal subjects. Michie PT, Fox AM, Ward PB, Catts SV, McConaghy N (1990): ERP indices of selective attention and cortical lateralization in schizophrenia. Psychophysiology 27:209-227. Molfese DL, Papanicolaou A, Hess TM, Molfese VJ (1979): Neuroolectrical correlates of semantic processes, in Begleiter H (ed), Evoked Brain Potentials and Behavior. New York: Plenum, pp 89-106. Nuechterlein KH, Dawson ME (1984): Information processing and attentional functioning in the developmental course of schizophrenk disorders. Schizophr Bull 10:160-203. Pfefferbaum A, Wenegrat BG, Ford JM, Roth WT, Kopell BS (1984): Clinical application of the P3 component of event-related potentials. II Dementia, depression and schizophrenia. Electroencephalogr Ciin Neurophysiol 59:104-124. Polich J (1986): Attention, probability, and task demands as determinants of P300 latency from auditory stimuli. Electroencephalogr Clin Neurophysio! 63:25 !-259. Polich J, Burns T (1987): P300 from identical twins. Neuropsychologia 25:299-304. Pritchard WS (1981): Psychophysiology of P300. Psychol Bull 89:506-540. P:i~chard WS (1986): Cognitive event-related potential correlates of schizophrenia. Psychol Bull 100:43-66. Rapaport D (1945): Diagnostic Psychological Testing. Chicago: Year Book. Rosman B, Wild C, Ricci J, Fleck S, Lidz T (1964): Thought disorders in the parents of schizophrenic patients: A further study utilizing the Object Sorting Test. J Psychiatr Res 2:2~ !-221. Saitoh O, Niwa S, Hiramatsu K, Kameyama T, Rymar K, Itoh K (1984): Abnormalities in late positive components of event-related potentials may reflect a genetic predisposition to schizophrenia. Biol Psychiatry 19:293-303. Schreiber H, Stolz-Born G, Rothmeier J, Kornhuber A, Kornhuber HH, Born J (1991): Endogenous event-related brain potentials and psychometric performance in children at risk for schizophrenia. Biol Psychiatry 30:177-189. Tucker PK, Rothwell SJ, Armstrong MS, McConaghy N (1982): Creativity, divergent and allusive thinking in students and visual artists. Psychol Med 12:835-841.
BIOLPSYCHIATRY
P.B. Ward et al
1992;31:650-660
Ward PB, CaRs SV, Armstrong MA, McConaghy N (1984): P300 and psychiatric vulnerability in university students. Ann NY Acad Sci 425:645--652. Ward PB, Catts SV, Fox AM, Michie PT, McConaghy N (1991a): Audi:o~, selective attention and event-related potentials in schizophrenia. Br J Psychiatry 158:534--539. Ward PB, McConaghy N, CaRs SV (1991b): Word association and measures of psychosis proneness in university students. Person Indiv Diff 12:473-480.