Occurrence and phenomenology of hallucinations in the general population: A large online survey

Although epidemiological studies report that hallucinations occur in 6–15% of the general population, little is known about their phenomenology. To overcome this paucity, this study investigates the phenomenological characteristics of hallucinations in the general population, by using a nationally promoted online survey to assess hallucination phenomenology in four sensory modalities, through a self-report version of the Questionnaire for Psychotic Experiences (QPE), in 10,448 participants (aged 14–88 years). The phenomenology of hallucinations was assessed if hallucinations reportedly occurred in the past month. In the past month, auditory hallucinations were reported most frequently (29.5%), followed by visual (21.5%), tactile (19.9%), and olfactory hallucinations (17.3%); hallucinations in two or more modalities were reported by 47.6%. Substantial numbers of participants rated their hallucinations as severe, due to negative content (16.0–31.6%), previous bothersome experiences (14.8–20.2%), ensuing distress (10.5–16.8%), and/or ensuing disfunctioning (12.7–17.3%). Decreased insight was found in 10.2–11.4%. Hypnagogia was reported by 9.0–10.6%, and bereavement hallucinations by 2.8%. Despite a low prevalence of delusions (7.0%), these phenomena were significantly associated with recent hallucinations, observed in up to 13.4% of the participants with hallucinations during the past week (p < 0.001). Our results indicate a wide variety of the phenomenology of hallucinations in the general population and support the existence of a phenomenological continuum.


Overall survey structure
Overall, the survey consisted of four parts (referred to as 1. main study, 2. signal recognition task, 3. optional questions, 4. hearing test), each of which was connected with a separate database. Two parts of the total survey (main study; optional questions) were part of the internal structure of the study website. The two other parts (signal recognition task; hearing test) relied on data collection from external online sources. All data from completed survey entries were saved within secured online databases that could only be accessed through involved study personnel. Incomplete entries were not registered. Supplementary figure 1a shows the complete algorithm that a participant was able to follow once entering the survey.
As shown in supplementary figure 1a, the first step for every participant was to enter the main study, and to provide informed consent and several demographic characteristics.
Participants younger than 14 years were directed to the 'kiddy' part of the main study, which contained questionnaires on hallucination-like experiences and imaginary friends, which were specifically designed to suit younger participants and are not included in the current study. Participants aged 14 and over were redirected to complete the complete main part of the study, consisting of questionnaires on hallucination-like experiences, sleep, and recreational drug use.
At the end of the main study the participant was alerted about completing the first part of the survey, and discretely allocated with an entry-specific research number. This concluding webpage also contained separate links inviting participants to continue their participation in two subsequent parts of the study: the signal recognition task (accessed through https://wvdw.coolminds.nl ) and the optional questions. Once accessed, the allocated research number was automatically entered in both databases, invisible to the participant, so that the outcomes of both surveys could be linked with data from the main one.
Within the optional questions section, participants were again questioned whether they were aged 14 and over or younger, on which the arrangement of the remaining questionnaires was based. Participants aged 14 and over were directed to questionnaires about atopic disease, delusion-like experiences, loneliness and augmented reality gaming.
Participants younger than 14 were only presented with questionnaires on atopic disease and augmented reality gaming.
Finally, all participants that completed the optional question part of the survey were presented with a link to the final part: the hearing test, which could be accessed through an    Fig.3i, 3j; total percentage exceeds 100%). The total amount of surveyed participants is 3,086 for every item, unless specified otherwise. The age of onset ( Figure 1d) has been computed by subtracting the answers from figure 2c from the participants' age. Due to an inaccurately programmed answer in the online version of QPE item 1.7 ("Repetition of AH"), this item has been excluded from current analysis in order to retain overall validity. In Figure 1o, a small sample of participants (n=55; 1.8%) provided answer combinations that appeared contradictory (i.e., both a specific and non-specific location). Similarly, in Figure  1p, 332 participants (10.8%) provided potentially contradictory answer combinations (i.e., 'non-verbal only' with one or more verbal answers). In both items, these entries have been maintained to provide the most accurate representation of the obtained data.
Supplementary Figure 4 Supplementary Figure 4. Additional phenomenological items on visual hallucinations (VH) (n=2,248), obtained through multiple choice items with one answer possibility (Supp. Fig. 4a-i) and multiple answer possibilities (Supp. Fig. 4j, 4k; total percentage exceeds 100%). The total amount of surveyed participants is 2,248 for every item, unless specified otherwise. The age of onset (Supp. Fig. 4h) has been computed by subtracting the answers from Supp. Fig. 4g from the participants' age. In Supp. Fig. 4j, a sample of participants (n=267; 11.9%) provided answer combinations that appeared contradictory (e.g., mostly left and mostly right). These entries have been maintained to provide the most accurate representation of the obtained data. 14 Supplementary