Psychiatry and Clinical Neurosciences

(2004),

58

, 651–659

Blackwell Science, LtdOxford, UKPCNPsychiatry and Clinical Neurosciences1323-13162004 Blackwell Science Pty Ltd586651659Original Article

Auditory hallucination phenomenologyN. Hayashi et al.

Correspondence address: Naoki Hayashi, Department of Psychiatry,Tokyo Metropolitan Matsuzawa Hospital, 2-1-1 Kamikitazawa,Setagaya-ku, Tokyo 156-0057, Japan. Email: nhayashi55@nifty.com

Received 16 March 2004; revised 6 May 2004; accepted 16 May2004.

Regular Article

Phenomenological features of auditory hallucinations and their symptomatological relevance

NAOKI HAYASHI,

MD, p

h

d

,

1,2

YOSHITO IGARASHI,

md

,

3

KIYOKO SUDA,

md

4

AND SEISHU NAKAGAWA,

md

1

1

Department of Psychiatry, Tokyo Metropolitan Matsuzawa Hospital,

2

Department of Mental Health Services, Tokyo Institute of Psychiatry,

3

Department of Psychiatry, Tokyo Metropolitan Fuchu Hospital, Tokyo, Japan and

4

National Centre for Register-based Research, Aarhus University, Aarhus, Denmark

Abstract

Auditory hallucinations include particularly diverse phenomena that reflect various mental func-tions and pathologies. Their assessment may provide valuable clinical information. This articledescribes the development of the Matsuzawa Assessment Schedule for Auditory Hallucination(MASAH), which was designed to obtain a broadened view of the phenomena by investigating awide range of their characteristics. The aim was to identify the basic phenomenological featuresof auditory hallucinations by performing a factor analytic study of the MASAH ratings of 214patients with schizophrenia or schizoaffective disorder. Four identified factors were intractability,delusion, influence, and externality, on the basis of which we constructed composite scales thatwere assumed to represent the features. The correlation analysis of the scales with symptom dimen-sions derived from the positive and negative syndrome scale verified their clinical relevance. Theywere also interpretable in terms of human responses to the abnormal experience and some symp-tom constructs such as delusion and influence experience. It is concluded that the MASAH is anefficient means for evaluating the features, and that this study elicited new understandings of thephenomena such as their multifarious composition and contiguities with other psychotic symptoms.

Key words

auditory hallucination, factor analysis, psychiatric rating scale, psychotic symptoms, schizophrenia.

INTRODUCTION

Auditory hallucinations are known to be a principalfeature of psychoses, and are especially commonamong patients with schizophrenia (e.g. Goodwin

et al

.

1

). They are also a potential source of distress forthe patients and, therefore, of clinical importance.Although the symptom is customarily classified as anexplicit perception abnormality, the actual experiencesare, more often than not, seen as end-products of per-ceptual and various other mental processes such ascognition and judgments, and reflect those functions ofthe patients. It is hoped that investigations into thephenomena may reveal their underlying pathology andimpact on mental activities.

The phenomenology of this symptom has repeatedlybeen alleged to have clinical significance. For example,Schneider counted three related symptoms: hallucina-tion of running commentary, voices speaking of thepatient in the third person, and audible thoughts, asamong the first rank symptoms (FRS) for the diagnosisof schizophrenia according to his definition.

2

Alongwith diagnostic values, the characteristics may be clin-ically informative. One example is that commandhallucination or hallucination of derogatory contentwould be suggestive of concomitant violent or suicidalbehavior.

3,4

Another significant aspect of the phenom-enological investigation would reside in its possibleusefulness for elucidating the heterogeneity of audi-tory hallucination. Explorations into its generatingmechanisms have suggested that complicated factorsmay be involved. A phenomenological study mightgive some indications useful for distinguishing differ-ent types of hallucinations that have different generat-ing factors.

652 N. Hayashi

et al.

Since the significance of auditory hallucinations washighlighted, a number of schedules that aimed atassessing the phenomena have been developed.

5–13

Theinvestigated areas included perceptual characteristics,patient responses, judgments and beliefs. We shouldalso take particular note of another area the previousassessments rarely dealt with: hallucinatory voice-related symptoms such as audible thoughts; ego distur-bance experiences caused by the voices and variousdelusional elaborations of the experience. This multi-plicity of characteristics indicates that the experiencehas multiple facets. However, there is no availableassessment that contains all the characteristics of thosefacets together, even though they are all basic constit-uents of the experience. To broaden our view of thephenomena, it would be a promising approach toincorporate the various characteristics noted above inthe assessment.

This report firstly describes the development of theMatsuzawa assessment schedule for auditory halluci-nations (MASAH), which contains items across thefacets of the experience and attempts to elucidate itsbasic features. Next, the clinical relevance of the fea-tures is sought by examining their correlation coeffi-cients with symptom dimensions of the positive andnegative syndrome scale (PANSS). These attemptsshould shed some light on the question of where audi-tory hallucinations are to be situated in the psychoticsymptomatology.

METHODS

Development of the Matsuzawa assessment schedule for auditory hallucinations

Auditory hallucination characteristics, on the basis ofwhich original MASAH item inquiries were produced,were basically adopted from our previous study

7

and some from the previous auditory hallucinationassessments

5,8–10

(items adopted common with or fromother assessments are indicated in Note 3 of theAppendix). The original item set had 48 items, andwere administered in the present study. The next stagewas selection and preparation for the final item set ofthe MASAH. Items with inadequate reliability scoresand those that were non-contributory to a factor ana-lytic study were removed from the item set. Items thatwere theoretically related and strongly correlated(Pearson correlation coefficient

>

0.5) in a preliminaryanalysis were grouped to compose a single item. The 23item final version was formed and used in further anal-yses. A sample scale item and the item list of theMASAH final version, are shown in the Appendix. TheMASAH consists of three sections: (i) perceptual char-

acteristics; (ii) patient responses and judgments(related or accompanying symptoms); and (iii) beliefsand general judgments. There are also two kinds ofitems in the MASAH: a total of 17 patient self-reportitems and six rater judgment items as indicated in Note2 of the Appendix. The English translations of theMASAH original and final versions are available onrequest to the first author.

Subjects

Subjects were enrolled by asking psychiatrists workingin Tokyo Metropolitan Matsuzawa Hospital, Japan tolist candidate patients for this study. Criteria for inclu-sion were: (i) currently experiencing or having recentlyexperienced auditory hallucination; (ii) diagnosis ofschizophrenia or schizoaffective disorder as per the4th edition of the Diagnostic and Statistical Manual(DSM-IV); (iii) not having prominent organic cogni-tive disorder or mental retardation; (iv) giving consentto participate in this study; and (v) being clinicallyjudged to be stable enough to undergo the assessment.A total of 214 patients participated in this study. Weobtained written informed consent from the subjectsfor their participation. The diagnosis was determinedon the basis of examining case records and additionaldiagnostic inquiries in the investigation.

Administration of instruments

According to the prepared protocol, the MASAHinterview was conducted to assess auditory halluci-nations experienced in the previous 2 weeks. Theinterviewer paid meticulous attention to not arousingdeleterious anxiety in the subjects during the interview.At the end of the MASAH, the interviewer attemptedto make the interview have some therapeutic meaningby providing the subjects with psychological support tomitigate their anxiety and by suggesting their use ofsome technique to cope with the experience.

Test-retest and interrater reliability studies of theMASAH items were conducted in a subgroup of 20currently hallucinating inpatients (six males and 14females) with an average age of 53.3 years. To examinetest-retest reliability, we interviewed the subgroup sub-jects twice, 7–10 days apart. In the interrater reliabilitystudy of the rater judgment items of the MASAH, tworaters made independent ratings of the scales on thebasis of a verbatim record of the MASAH interview.

In another major subgroup of 141 subjects who werecurrently hallucinating, clinical symptoms were ratedaccording to PANSS. Our research group had held atraining course for reliable PANSS rating skills by uti-lizing six video-recorded interviews. Four raters in the

Auditory hallucination phenomenology 653

present study fulfilled the criteria of Kay

et al

.

14

by theend of the training, that is, the concordance rate for thewhole set of items (a deviance of one point was alsoincluded as concordant) must be over 80% and scoresof each subscale must be within the range of 80–120%of consensus ratings.

Data analysis

In the reliability studies of the MASAH items, the

anova

intraclass correlation coefficient (

anova

ICC)was calculated. The items that were confirmed to havea significant ICC (

anova

ICC

>

0.40) were included infurther analyses. Next, principal-components analysiswith varimax rotation of the items was performed toextract basic features of auditory hallucinations. Thenumber of factors for the factor analysis was decidedby inspection of a scree plot of the factors. In addition,we attempted to attain a simple factor structure bysequentially eliminating the items not having absolutevalues of its factor loadings greater than 0.5 (not beinga principal component of any factor). Subsequently,composite scales of the extracted factors were con-structed by adding or subtracting scores of principalcomponents of each of the factors according to thesigns of the factor loadings. Lastly, we conducted cor-relation analysis of the composite scales with demo-graphic variables and five symptom dimensionsderived from PANSS to explore the clinical signifi-cance of the features that the composite scales wereassumed to represent. Although the PANSS symptomfactor model has not been established, our review

15

found that the previous major factor analytic studies ofPANSS that treated with more than 200 subjects con-

sistently supported the five-factor model. Therefore,the present study applied the five symptom dimensionmodel introduced by Kay

et al

.

16

which is composed of‘anergia’, ‘thought disturbance’, ‘activation’, ‘paranoid/belligerence’, and ‘depression’. For the entire dataanalysis, the SPSS (Release 10.0.0; SPSS, Chicago, IL,USA) statistical software package was used.

RESULTS

The demographic and clinical data for the subjects areshown in Table 1. As indicated in the table, a noticeablecharacteristic of the subjects was that the course ofillness and hallucinatory experience were chronic formost of the subjects.

Reliability study

anova

ICC calculated to examine the test-retest reli-ability of the original MASAH items verified the reli-ability for the most part. A total of 94% of the originalMASAH items (45/48) showed significant reliabilityvalues (

anova

ICC

>

0.40). The average

anova

ICC(SD) for the MASAH final version items was 0.82(0.12). The interrater reliability for six rater judgmentitems proved to be good to excellent; their average

anova

ICC (SD) was 0.82 (0.15).

Factor analytic study of the characteristics

A four-factor solution was decided by inspection of thescree plot of the preliminary factor analysis of allvariables with acceptable reliability, and found to betheoretically meaningful. Table 2 presents the factor

Table 1.

Demographic and clinical characteristics of the subjects

Total subjects (

n

=

214) Subgroup subjects (

n

=

141)

Male/Female (%) 116 (54)/98 (46) 61 (43)/80 (57)Inpatients/Outpatients (%) 179 (84)/35 (16) 110 (78)/31 (22)Diagnosis (Schizophrenia/Schizoaffective disorder) 202/12, 94%/6% 133/8, 94%/6%Shorter than 3 years duration of auditory hallucination 91, 43% 34, 24%Use of atypical antipsychotics

†

64, 30% 43, 30%Age at investigation (years) 45.0 (14.8), 19–75 50.3 (14.1), 19–75Age at onset (years) 27.4 (9.8), 10–64 26.9 (9.9), 10–64Education (years) 12.0 (2.3), 6–16 11.7 (2.3), 6–16Lifetime hospitalizations 3.4 (3.5), 0–28 4.1 (4.0), 0–28Positive Subscale (PANSS) 42.9 (4.9), 13–35Negative Subscale (PANSS) 23.4 (5.9), 11–37General psychopathology (PANSS) 42.7 (8.4), 25–65

†

Values below this are expressed as means (SD), range.PANSS, positive and negative syndrome scale.

654 N. Hayashi

et al.

structure attained in the factor analysis of the 23MASAH items. The variances explained by prerota-tional factors were 19.4%, 14.9%, 10.2%, 7.0%, 5.4%and 4.6% in order of extraction.

The first factor had six principal components (itemswith factor loadings greater than 0.50) that indicatedthe intractable nature of the experience, and wasnamed intractability. The components were negativevoice content, negative patient responses (emotion,judgments and beliefs), and uncontrollability of thevoices. Next, the components of the second factor,delusion; generalization of a delusional theme, unreal-istic specifications, loss of symptomatic insight andunproportional conviction, appeared to be cut-in ele-ments of delusional reality distortion. The third factor,influence, included excessive influence of the voicesand the patient’s precarious subjectivity in the experi-ence. The fourth factor, externality, was composed offive components of perception styles or hearing-voices

situations that appeared to indicate external or internallocalization of the voices and their origins.

Correlation coefficients between the composite scalescores and regression factor scores, and Cronbach’s

a

coefficients and

anova

ICC for testing test-retest reli-ability of the scales are also presented in Table 2. Thescale scores were highly correlated with their corre-sponding factor scores. Therefore, we gave the scalesthe same names as their factors. Three of the four com-posite scales showed permissable internal consistency(Cronbach’s

a

>

0.7). The test-retest reliability of thescales proved to be excellent as shown in the last line ofthe table. The correlations between two of the scalescores were modest, ranging between 0.136 and 0.278.

Correlation analysis of the composite scales

The results of correlation analysis of the compositescales with PANSS symptom dimensions are presented

Table 2.

The factor structure for the Matsuzawa Assessment Schedule for Auditory Hallucination characteristics

F1: Intractability F2: Delusion F3: Influence F4: Externality

Unpleasant feelings 2-(1) 0.838

†

Hostility of voices 3-(4) 0.794

†

Malevolent content 1-(1) 0.789

†

Distress 2-(2) 0.745

†

0.397Recognition of failure in coping 3-(1) 0.658

†

Controllability of voices 1-(7)

-

0.617

†

Generalization of accompanying delusion 2-(7) 0.758

†

Conviction of delusion 2-(8) 0.734

†

Unusualness of origin 1-(4)

-

0.366 0.634

†

Delusional explanation 2-(6) 0.593

†

0.314Attribution of voices to illness 3-(2)

-

0.573

†

Identification of origins 1-(3) 0.561

†

Actualness 3-(3) 0.549

†

Ego disturbance in the inner world 2-(5) 0.742

†

Audible thoughts 2-(4) 0.607

†

Influence of voices 1-(8) 0.606

†

Preoccupation 2-(3) 0.415 0.587

†

Imperative content 1-(2) 0.527

†

-

0.411Voices directly speaking to patient 1-(9)

-

0.697

†

Conversation among voices 1-(10) 0.323 0.657

†

Outside location of origins 1-(5) 0.634

†

Perception through ears 1-(6) 0.587

†

Voices from present figures outside 1-(11) 0.522

†

Pearson correlation coefficient

‡

0.968 0.976 0.922 0.961Cronbach’s

a

§

0.856 0.762 0.813 0.643

anova

ICC (Test-retest reliability) 0.950 0.852 0.899 0.930

Only items that loaded 0.30 or greater are shown in this table.

†

Indicates principal components (factor loading

>

0.50);

‡

The correlation coefficient between the factor’s composite scale scoreand regression factor score;

§

Constituent items of principal components are also included in the calculation.ICC, intraclass correlation coefficient.

Auditory hallucination phenomenology 655

in Table 3. The correlation coefficients appeared to beconsistent with connotations the scale names wouldhave. ‘Thought disturbance’ had mild to moderate cor-relations with ‘delusion’, ‘influence’ and ‘intractability’.‘Paranoid/belligerence’ showed slight but significantcorrelations with ‘externality’ and ‘delusion’. ‘Depres-sion’ mildly correlated with ‘intractability’ and ‘influ-ence’. In contrast, ‘anergia’ and ‘activation’ showedonly minor correlations with the features.

The correlation analysis of the feature scales withdemographic and clinical variables contained inTable 1 revealed significant Spearman’s correlationcoefficients between ‘delusion’ score and age at inves-tigation (0.334,

P

<

0.001), duration of illness (0.261,

P

<

0.001), and duration of the hallucinatory experi-ence (0.237,

P

=

0.001). Since these clinical variableswere intercorrelated, partial correlation analyses werealso conducted. The correlation of age at investigationwith the delusion feature remained significant whenthe effects of the other two variables were removed(0.243,

P

<

0.001, one-tailed), while the correlations ofthe other two became non-significant when controlledfor the effect of age at investigation. Therefore, a linkbetween older age and the delusional feature was mostlikely.

DISCUSSION

Clinical significance of phenomenological features

By virtue of detailed inquiries of the MASAH, thepresent study identified the factors that would repre-sent clinically interpretable phenomenological featuresof the experience, on the basis of which we constructedscales that can be used for further investigations.

The intractability feature concerns the patient’sbasic attitude to the experience; whether the patientfinds the experience acceptable and controllable ornot, and involves the emotional, judgmental and

behavioral responses of the patient to the experience.Its inverse relationship with the thought disorderdimension of PANSS could be explained by thepatient’s consistent attitude that this feature implies.This is thought to be closely related to the malevolencefeature and inversely related to the benevolence fea-ture of auditory hallucination defined in the study ofChadwick

et al

.

13

Those authors also verified that thesebeliefs were correlated with depressive symptoms, aswas found in the present study with respect to theintractability feature.

The characteristics of delusion ingrained in auditoryhallucinations and the delusion-related cognition andjudgments were grouped in the delusion feature in thepresent study. This finding ascertained that the delu-sional characteristics would be inherent in auditoryhallucinations. This feature is thought to be continuouswith definite delusion and gives some explanation forthe clinically acknowledged interplay between thehallucinatory experience and delusion. This study alsoconfirms the relation between delusion and thoughtdisorder, which has been considered to be intrinsic (e.g.Harrow and Quinlan

17

). In addition, the present studysuggested that this feature could be intensified withincreasing age. This may be a manifestation of the rela-tionship between delusion-proneness and aging amongpatients with schizophrenia that was reported byHäfner

et al

.

18

The factor analysis of this study grouped symptomsof voice-related ego disturbance in the inner world

2,19

and those indicating excessive influence by the voicesinto the influence feature. They are suggestive of thepatient’s condition of weakened subjectivity. It isjustifiable that ‘audible thoughts’ was included herebecause this abnormal self-attribution of the patient’sown thoughts indicates a precarious condition of sub-jectivity. Most such symptoms are included in thenuclear schizophrenia symptoms of the present stateexamination (PSE)

20

and the Schneider’s FRS, and,therefore, are regarded as core symptoms of schizo-

Table 3.

Correlation analysis of the Matsuzawa Assessment Schedule for Auditory Hallucination composite scales with positiveand negative syndrome scale symptom dimensions

Intractability Delusion Influence Externality

Anergia 0.180Thought disturbance

-

0.297

‡

0.470

‡

0.275

‡

Activation 0.203Paranoid/belligerence 0.209 0.221

†

Depression 0.336

‡

0.283

‡

0.176

Only significant Pearson correlation coefficients (

P

<

0.05, two-tailed) are shown in this table.

†

Correlation is significant at the 0.01 level;

‡

Correlation is significant at the 0.001 level.

656 N. Hayashi

et al.

phrenia. The feature components also comprise a partof the bizarre delusions defined in DSM-IV, whichmight explain its correlation with the thought disorderdimension of PANSS. This feature also seems to corre-spond to the voices’ omnipotence in the study ofChadwick

et al

.

13

Those authors stressed the clinicallyimportant fact that this feature would determine thepatient’s relationship with the voices, and reported itspositive correlation with depressive symptoms, whichis comparable to the findings of this study.

The feature of externality involves the patient’s per-ception styles or experiencing situations that indicateexternal or internal localization of hallucinatory voices.For instance, ‘voices directly speaking to patients’ indi-cates an imminent and distance-loss nature of theexperience, that is, internal experience location. Theexternal localization may be a part of paranoid symp-toms that stress excessive hostility and suspicion toouter objects. The correlation between the externalityand the paranoid symptom dimension found in thisstudy supports such a relationship. This feature is alsorelated to the pathology of sound source localization,and may have some neurocognitive bases, which sev-eral studies have attempted to identify experimen-tally,

21–23

and may be a factor in determining the stylesof the experience.

Thus, the features identified in the present study areinterpretable in terms of some symptom constructs andhuman responses to the experience, and would haveclinical and research relevance.

Except for our previous study,

7

only the study ofSingh

et al

.

24

conducted a factor analytic study thatfocused on auditory hallucination phenomenology,which yielded two factors: reality of hallucinatory per-ception and immersion in hallucination. It is quite dif-ficult to compare this study with that of Singh

et al

.

24

since this study dealt with the phenomena differentlyand more comprehensively. However, it is noticedthat the factor of reality of hallucinatory perceptionappeared to share delusional reality distortion in thephenomena with the delusion factor of this study.

Psychotic symptomatology and auditory hallucination

In a general frame of reference, auditory hallucinationis classified as the major feature of the positive syn-drome along with delusion, which many empiric stud-ies have verified (e.g. Liddle25 and Andreasen et al.26).However, a closer look reveals different views of amore complicated symptom structure. A factor analyticstudy of PSE data by Mellers et al.27 found that the pos-itive syndrome could be divided into two factors: audi-tory hallucination factor and delusion factor. Based on

the scales for the assessment of positive and negativesymptoms (SANS and SAPS) investigations, Cuestaand Peralta28 and Stuart et al.29 identified an auditoryhallucination factor apart from delusion factors.Cuesta and Peralta30 also showed that auditory hallu-cination items of AMDP (Arbeitsgemeinschaft fürMethodik und Dokumentation in der Psychiatrie Sys-tem) pertained to the Schneider’s FRS factor otherthan delusion factors. Similarly, the study of diagnosticvalues of psychotic symptoms by Wing and Nixon31 alsoplaced ‘voices directly speaking to the patient’ near theSchneider’s FRS. The factor analytic study of Kitamuraet al.32 suggested that there are two groups of auditoryhallucinations: ‘voices directly speaking to the patient’,‘conversation among voices’, ‘voices coming frominside the body’ and ‘running commentary’ that wereincluded in the Schneider’s FRS factor, and other audi-tory hallucinations that were grouped in the hallucina-tion factor together with visual hallucination. Takentogether, these findings indicate that auditory halluci-nations are subtly distinct from delusion symptoms,and that some of them are situated in or close toSchneider’s FRS. This variable affiliation pattern ofauditory hallucinations also indicates the need to scru-tinize the heterogeneity of auditory hallucinations andto locate properly their positions in psychotic symp-tomatology. For those purposes, the present studywould be a novel contribution. This study has proposeda finer symptom structure of auditory hallucinationsand demonstrated their contiguities with other symp-tom constructs, such as delusion and influence ex-periences including the Schneider’s FRS or nuclearschizophrenia symptoms, which may explain to someextent some discrepancies among the previousresearch findings about auditory hallucinations.

It is necessary to mention some weak points of thisstudy that should be remedied in future research. First,the subjects of this study might not be representative ofpatients who experience auditory hallucinations sincethe subjects were shifted to chronic and severe cases. Itshould also be noted that the sample was not formed ina systematic sampling procedure. Second, the subjectsincluded patients in various phases of their illness.Phase-dependent characteristics that could not be dis-cerned in this study might have affected the results.Longitudinal studies that treat patients in the samestage of illness can remedy these weaknesses. Last, theselection and categorization of the MASAH items arestill tentative, and require elaboration with the helpof future advances in understanding this abnormalexperience.

Despite the limitations, we may draw some conclu-sions from the present study. First, the results appear toconfirm that the MASAH would be an efficient means

Auditory hallucination phenomenology 657

for assessing the experience of auditory hallucinations.Next, the features derived from the MASAH wouldclinically be usable for understanding the patients hav-ing the experience. They might also be a potent meansfor investigating the heterogeneity and underlyingpathologies of the experience. Consequently, thepresent investigation suggests the potential value offurther research on the significance of auditory hallu-cinatory phenomena in psychiatric practice and psy-chotic symptomatology.

ACKNOWLEDGMENTS

The authors wish to thank Drs Ryosuke Nakamura,Naoko Ishige, Yukiyo Inoue, Hirohiko Harima, Hide-masa Onai, Shusuke Yoneda, Taiki Tao and YuichiYamashita at Tokyo Metropolitan Matsuzawa Hospitalfor cooperation in this study. This study was supportedby a Grant-in-aid for exploratory Research (no.11877164) from the Japan Society for the Promotion ofScience.

REFERENCES

1. Goodwin DW, Alderson P, Rosenthal R. Clinical signif-icance of hallucinations in psychiatric disorders: A studyof 116 hallucinatory patients. Arch. General Psychiatry1971; 24: 76–80.

2. Schneider K. Klinische Psychopathologie (Clinical Psy-chopathology). Georg Thieme, Stuttgart, 1962.

3. Zisook S, Byrd D, Kuck J, Jeste DV. Command halluci-nations in outpatients with schizophrenia. J. Clin. Psychi-atry 1995; 56: 462–465.

4. Cheung P, Schweitzer I, Crowley K, Tuckwell V. Vio-lence in schizophrenia: role of hallucinations and delu-sions. Schizophr. Res. 1997; 26: 181–190.

5. Lowe GR. The phenomenology of hallucinations as anaid to differential diagnosis. Br. J. Psychiatry 1973; 123:621–633.

6. Junginger J, Frame CL. Self-report of the frequency andphenomenology of verbal hallucinations. J. Nerv. Ment.Dis. 1985; 173: 149–155.

7. Hayashi N. Analysis of Auditory Hallucinations of Psy-chotic Patients: Multivariate Analysis Approach I. Seish-inigaku 1986; 27: 267–278 (in Japanese).

8. Hustig HH, Hafner RJ. Persistent auditory hallucinationand their relationship to delusion and mood. J. Nerv.Ment. Dis. 1990; 178: 264–267.

9. Oulis PG, Mavreas VG, Mamounas JM, Stefanis CN.Clinical characteristics of auditory hallucinations. ActaPsychiatr. Scand. 1995; 92: 97–102.

10. Carter DM, Mackinnon A, Howard S, Zeegers T,Copolov DL. The development and reliability of theMental Health Research Institute Unusual PerceptionsSchedule (MUPS): an instrument to record auditory hal-lucinatory experience. Schizophr. Res. 1995; 16: 157–165.

11. Nayani TH, David AS. The auditory hallucination: aphenomenological survey. Psychol. Med. 1996; 26: 177–189.

12. Haddock G, McCarron J, Tarrier N, Faragher EB. Scalesto measure dimensions of hallucinations and delusions:the psychotic symptom rating scales (PSYRATS). Psy-chol. Med. 1999; 29: 879–889.

13. Chadwick P, Lees S, Birchwood M. The revised beliefsabout voices questionnaire (BAVQ-R.). Br. J. Psychiatry2000; 177: 229–232.

14. Kay SR, Opler LA, Lindenmayer JP. Reliability andvalidity of the positive and negative syndrome scale forschizophrenics. Psychiatry Res. 1988; 23: 99–110.

15. Hayashi N, Igarashi Y, Yamashina M, Suda K. Is theregender difference in a factorial structure of the Positiveand Negative Syndrome Scale?: a test by structural equa-tion modeling. Psychopathol. 2002; 35: 28–35.

16. Kay SR, Opler LA, Fiszbein A, Ramirez PM, White LW.Positive and negative syndrome scale (PANSS) user’smanual. Multi-Health Systems, Toronto, 2000.

17. Harrow M, Quinlan DM. Disordered Thinking andSchizophrenic Psychopathology. Gardner Press, NewYork, 1985.

18. Häfner H, Löffler W, Riecher-Rössler A, Häfner-Ranabauer W. Schizophrenie und Wahn im höherenund hohen Lebensalter. Nervenarzt 2001; 72: 347–357 (inGerman).

19. Jaspers K. Allgemeiner Psychopathologie, 5th edn.Springer, Berlin, 1948.

20. Wing JK, Cooper JE, Sartorius N. Measurement andClassification of Psychiatric Symptoms: An InstructionManual for the PSE and Catego Program. CambridgeUniversity Press, Cambridge, 1974.

21. Balogh DW, Schuck JR, Leventhal DB. A study ofschizophrenics’ ability to localize the source of a sound.J. Nerv. Ment. Dis. 1979; 167: 484–487.

22. Heilbrun AB Jr, Blum N, Haas M. Cognitive vulnerabil-ity to auditory hallucination: Preferred imagery modeand spatial location of sounds. Br. J. Psychiatry 1983; 143:294–299.

23. Keefe RS, Arnold MC, Bayen UJ, Harvey PD. Sourcemonitoring deficits in patients with schizophrenia: a mul-tinominal modeling analysis. Psychol. Med. 1999; 29:903–914.

24. Singh G, Sharan P, Kulhara P. Phenomenology of hallu-cinations: a factor analytic approach. Psychiatry Clin.Neurosci. 2003; 57: 333–336.

25. Liddle PF. The symptoms of chronic schizophrenia: a re-examination of the positive-negative dichotomy. Br. J.Psychiatry 1987; 151: 145–151.

26. Andreasen NC, Arndt S, Alliger R, Miller D, Flaum M.Symptoms of schizophrenia: Methods, Meanings, Mech-anisms. Arch. General Psychiatry 1995; 52: 341–351.

27. Mellers JD, Sham P, Jones PB, Toone BK, Murray RM.A factor analytic study of symptoms in acute schizophre-nia. Act. Psychiatr. Scand. 1996; 93: 92–98.

28. Cuesta MJ, Peralta V. Dimensional structure of psy-chotic symptoms: an item-level analysis of SAPS and

658 N. Hayashi et al.

SANS symptoms in psychotic disorders. Schizophr. Res.1999; 38: 13–26.

29. Stuart GW, Pantelis C, Klimidis S, Minas IH. The threesyndrome model of schizophrenia: Meta-analysis of anartefact. Schizophr. Res. 1999; 39: 233–242.

30. Cuesta MJ, Peralta V. Integrating psychopathologicaldimensions in functional psychoses: a hierarchicalapproach. Schizophr. Res. 2001; 52: 215–229.

31. Wing J, Nixon J. Discriminating symptoms in schizophre-nia: A report from the international pilot study of schizo-phrenia. Arch. General Psychiatry 1975; 32: 853–859.

32. Kitamura T, Okazaki Y, Fujinawa A, Takayanagi I,Kasahara Y. Dimensions of schizophrenic positive symp-toms: an exploratory factor analysis investigation. Euro.Arch. Psychiatry Clin. Neurosci. 1998; 248: 130–135.

Appendix 1. The (a) Matsuzawa Assessment Schedule for Auditory Hallucination (MASAH) sample scale and the (b) list ofthe MASAH items

(a) Sample scale – Section 1 (1) Malevolent (benevolent) contentWhat are the contents of (the experience in the patient’s words such as unusual voices, rumors or whispers)?Are their contents benevolent, malevolent or neutral?

Check or fill in the following items and brackets:Benevolent content, • helpful; • assisting; • encouraging; • affirmative; • praising; • other ( )Malevolent content, • accusatory; • critical; • threatening; • despising; • other ( )Neutral content, ( )

Choose the most suitable for your experience:1. Totally benevolent content2. Mainly benevolent content3. Balanced or neutral or can’t answer4. Mainly malevolent content5. Totally malevolent content

(b) The list of the MASAH itemsSection 1. Perceptual characteristics (rated on five-point scale except items (7) and (8))Contents

(1) malevolent (benevolent) content (benevolent–malevolent)(2) amount of imperative content (never–mostly)

Nature of origins and perceptual styles(3)† identification of the origins (unclear–identifiable)(4)† unusualness of the origins (realistic–totally absurd)(5) outside location of the origins (inside–outside)(6) perception through ears (through other body parts–through ears)(7) controllability (the sum of three three-point [absent–unclear–present] scale scores of the patient’s capabilities [i] to start,

[ii] to stop voices and [iii] to change their tones or contents)(8) influence of the voices (the sum of scores of three three-point scales [absent–unclear–present] of the influence on the

patient’s [i] behavior, [ii] thoughts and [iii] feelings)Situation of hearing voices

(9) hearing voices directly speaking to the patient (never–mostly)(10) voices talking about the patient in the third person (conversation about the patient among voices; never–mostly)(11)† voices from actually present figures situated in the outside (recognizable by sight; never–mostly)

Section 2. Patient responses and judgments (related or accompanying symptoms; rated on a five-point scale except item (5))Subjective responses and judgments

(1) unpleasant feelings (pleasant–unpleasant)(2) distress (not distressing–distressing)(3) preoccupation (not preoccupied–preoccupied)

Related or accompanying symptoms(4) audible thoughts (never–mostly)(5) hallucinatory voice-related ego disturbance in the inner world (the sum of six three-point [absent–unclear–present] scale

scores of the voice-related [i] made experience [delusion of control], [ii] thought insertion, [iii] mind reading [delusion of thoughts being read], [iv] thought withdrawal, [v] thought broadcast, and [vi] delusion of being observed)

Auditory hallucination phenomenology 659

(6)† delusional explanation (electricity, telepathy, etc.; never–extended)(7)† generalization (no delusion or least–highly generalized)(8) conviction of accompanying delusion (unsure–completely sure)

Section 3. Patient beliefs and general judgments (rated on five-point scale)(1) recognition of success in coping (failure–success)(2) attribution of the voices to illness (symptomatic insight; due to illness–due to a reason other than illness)(3) actualness of the voices (imagery–actual)(4) friendliness of the voices (hostile–friendly)

Common items with previous assessments are as follows (asterisks indicate the items that were not included in our previous study7 and adopted from other studies): Items 1-(5), 1-(7) and 3-(2)* were included in Lowe;5 items 2-(2)* in Hustig and Hafner;8 items 1-(3), 1-(11), 2-(7) and 3-(3)* in Carter et al.;10 items 1-(3) and 1-(5) in Junginger and Frame;6 items 1-(3), 1-(5), 1-(6), 3-(2)*, 3-(3)* and 3-(4)* in Oulis et al.;9 items 1-(1), 1-(6), 1-(7) and 2-(2) in Haddock et al.;12 and items 1-(10) and 2-(4) in Schneider’s first rank symptoms for the diagnosis of schizophrenia.

Prior to inquiring about the items listed above, the interviewer is required to assess the frequency, clarity and loudness of the voices according to Hustig and Hafner8 and the duration of the experience.

†Indicate the rater judgment scales.

Appendix 1. Contintued