Characterizing Auditory Hallucinations:
An Aid in the Differential Diagnosis of Malingering

Julie A. Holland, M.D.  and Kevin C. Riley, PhD

        Since the famous paper by Rosenhan, "On Being Sane in Insane Places," (1) the concept of malingering psychosis has been an emotionally charged issue.  It puts our talents to the test, and we must rely on intuition and collateral sources to arrive at a diagnosis and  a safe disposition plan.  Although most of the literature on malingering focuses on the forensic population, in the psychiatric emergency room we also see patients who report the presence of auditory hallucinations in order to obtain food and shelter, to receive medication, or perhaps simply to enjoy the benefits of a therapeutic mileu. (2)  We are typically taught that the best teacher is experience, and over time we will get a sixth sense to assist us in detecting the patients who attempt to deceive us.  Since the phenomenon of auditory hallucinations is fairly well described in the psychiatric literature, one helpful strategy towards identifying malingerers, is to spend time with the patient reviewing the phenomenology of their "voices.."  The more we can learn about how auditory hallucinations are experienced by genuinely ill psychiatric patients, the better we can judge the veracity of a patient's subjective report.


        The following data was obtained by interviewing 86 psychiatric inpatients at a major metropolitan teaching hospital in Philadelphia, PA.  Patients from the psychiatric inpatient unit who had recently experienced auditory hallucinations were asked to participate as subjects.  Priority was given to actively hallucinating patients.  Patients were excluded whose primary language was not English, or whose Mini Mental Status Exam was less than twenty.  Sixty percent of approached patients agreed to be interviewed; informed consent was obtained after the experiment had been fully explained to the patient.  Subjects underwent a detailed structured interview to evaluate the form, content, frequency, extent, and intensity of hallucinatory experiences.


        Eighty seven subjects completed the structured interview, 39 women and 48 men.  Ages ranged form 17 to 75, with a mean of 36 years.  The average education of subjects was 11 years, ranging from 5to 18.  Sixty percent of subjects were either unemployed or unskilled laborers.  All subjects were medicated, most commonly with antipsychotics (85%) and/or benzodiazepines (41%).  Twenty percent of subjects were on antidepressant medications as well.  Urine drug screens were performed on roughtly half of the subjects as part of the admission work up, with 7 urines positive for cocaine, 1 for opiates, and 0 for methamphetamine.  Neuroradiologic studies were performed on 21 subjects.  Of the 17  CT's and 7 MRI's, five patients had abnormal findings: two had asymmetrical ventricals, one had frontal lobe atrophy, one had non-specific increased density signals in the right centrum semiovale by T2 weighted imaging, and one showed a basilar artery compression at the level of the foramen magnum.
        About one half of the interviewed group carried a diagnosis of schizophrenia, while the rest of the subjects carried a diagnosis of mood disorders or toxic/organic states, such as substance induced psychotic disorder or psychosis secondary to general medical condition.  The subjects were broken down into three categories in order to test some hypotheses about hallucinatory content and diagnostic specificity.  There was no difference between groups in any of the demographic data except age.  The toxic/organic group (N= 15)  had a mean age of 46 years, and was significantly older (p=.02) than the schizophrenic (N= 48) and mood disorder (N= 23) patients (mean ages 36 and 33 years, respectively).

Meaningfulness of content

        Ninety-nine percent of our sample heard some speech at one time.  Twenty percent indicated that at times they heard non-vocal stimuli such as noises, ringing, buzzing, music, or other.   Eighty-three percent of the sample could understand what was being said when hallucinations were verbal. While 7% of the sample heard commands, 48% overheard conversations regarding themselves, and 61% noted being called names.   Only 19% of the sample endorsed hearing a narrative of their actions.  Accusations (34%), threats (34%), and niceties (31%) were also reported.
        Of the patients who reported conversations, thirty-six percent  described the conversations as about them, while 23% said they were not about them, and 29% described the conversations as including them, in that they spoke to their voices and their voices spoke back to them.  Fifty-six percent said they recognized the voices: 30% friends, 29% family, 14% religious (angels, Jesus Christ, or God) and 13% occult.  Of note, no significant differences across our diagnostic groups in terms of content of the hallucinations was observed.

Characterization of Hallucinatory Experience

        Thirty eight percent of subjects surveyed described the sound as stereophonic, while six percent said it came from the right side, and four percent said left.  Ten percent ascribed the origin of sound as coming from behind them, and nine percent said from the front.  Thirty-three percent of interviewed subjects chose "other."
        Fifty-two percent of subjects could not ascribe a gender to the voice, and considered it to be a "mix," while 24% considered the voice to be male and 5% female. Five percent could not choose any category.  There was a non-statistical trend towards male subjects' hearing men's voices.
        The number of voices heard averaged 4.3, with no difference seen between the three groups.  When asked if the multiple voices spoke at the same time, 51% said different things were said at once, while 28% said the voices took turns speaking, and 11% said that multiple voices spoke in unison.
        When asked to choose between inside or outside of the head, the split is nearly equal, with 51% endorsing inside, 40% choosing outside, and 9% describing a combination of the two.  Seventy-nine percent reported the sound as being in both ears, with 4% right and 9% left ears.  Interestingly, when subjects were asked if they heard the same thing in each ear, 74% endorsed, while 26% denied this assumption.
        Eighty-five percent of subjects were right handed, 7% left and 8% mixed handedness.  Sixty one interviewed patients underwent dichotic listening tests of neuropsychological lateralization, and no consistent hemispheric deficits were seen within diagnostic groups.


        Auditory hallucinations are among the most prominent and distressing symptoms of a psychotic episode .  Form and content of hallucinatory experiences have been used as  significant clinical signs with diagnostic implications.  The textbooks teach us that hallucinations are generally experienced as originating in the outside world or within one's body, but not within the mind as through imagination.  It has been hypothesized that patients who view their voices as coming from inside the head may have a less distorted sense of reality, and they may be healthier if they are aware that it is their own thoughts they are hearing. (3)
        Goodwin undertook a study in 1971 which disproved various traditional assumptions about the clinical significance of hallucinations in psychiatric disorders.  Among these assumptions were that visual hallucinations signified organic brain syndrome, toxic hallucinations were experienced as outside the head, and nontoxic or endogenous hallucinations were considered by the patient to be inside.  Other diagnostic implications were that accusatory voices signified depression, commanding voices would lead to a dangerous act, and male schizophrenics heard male voices. (4)
        We were not able to draw any statistically significant correlations based on our data.  Because of my personal experience interviewing all 86 subjects, I feel that I can better assay whether a patient is truly hallucinating or not.  I teach psychiatric residents that they should never make a disposition plan based on the subjective data from the patient alone.  If the patient is exhibiting objective signs such as a formal thought disorder, delusions, or bizarre or blunted affect, these will help to support a label of psychotic.  When collateral  information is not available, direct observation of the patient when not being interviewed is often helpful, as is repeated interviewing.  Fatigue may help to create inconsistencies in the patient's story. (5) If one is assuming a diagnosis of major depression with psychotic features, note the patient's affect, appetite or presence of insomnia in the observation area of the ER.
        A urine toxicology screen can also help to clear up whether the disturbance is substance-induced, and frequently points the clinician towards a diagnosis of malingering in a patient who denied substance abuse prior to a positive result.  Other aspects of a rule-out malingering case to examine are: a presence of secondary gain, inconsistencies with prior presentations, and improbable or atypical symptoms or course.
        One important caveat for clinicians is that there is frequently true psychiatric illness in patients who occassionally present as malingering.  A 1967 paper by Braginsky and Braginsky (6) demonstrated that schizophrenics can minimize or exaggerate their known symptomatology in order to appear more or less ill, depending on their preference for disposition.  A 1983 study by Hay (7) found that out of six malingered psychoses, five eventually met criteria for a diagnosis of schizophrenia.  In these cases, a label of pseudomalingering would be more appropriate.


1.  Rosenhan, DL  "On Being Sane in Insane Places" Science, 1973 179: 250-258

2.  Yates BD, Nordquist CR, Schultz-Ross RA "Feigned Psychiatric Symptoms in the Emergency Room" Psychiatric Services 1996 47: 998-1000

3.  Asaad, G Hallucinations in Clinical Psychiatry, 1990 New York, Brunner and Mazel.

4.  Goodwin, D "Clinical Significance of Hallucinations in Psychiatric Disorders," Archives of General Psychiatry, 1971 24:76-80

5.  Resnick, Phillip "Malingered Psychosis" in Clinical Assessment of Malingering and Deception  Richard Rogers, editor, 1988 Guilford Press

6.  Braginsky BM , Braginsky DD "Schizophrenic patients in the psychiatric interview: an experimental study of their effectiveness at manipulation,"
Journal of  Consultation Psychology 1967 31:543-547

7.  Hay GG "Feigned psychosis--a review of the simulation of mental illness," British J ournal of Psychiatry 1983 143:8-10