What is pareidolia

Pareidolia is the interpretation of previously unseen and unrelated objects as familiar due to previous learning ¹. Pareidolias are visual illusions of meaningful objects such as faces and animals, are thought to arise from ambiguous forms embedded in visual scenes and have a striking phenomenological resemblance to visual hallucinations ². For example, patients with dementia with Lewy bodies may incorrectly see a person in a curtain or perceive blobs on the wall as faces.

Is pareidolia a mental illness?

It depends. Pareidolia is a phenomenon of recognizing patterns, shapes and familiar objects — often faces — where they do not actually exist ³. Faces convey primal information for your social life. This information is so primal that people sometimes find faces in non-face objects. Pareidolia is recognized in healthy humans as young as eight to 10 months of age ⁴.  Pareidolia has also been reported to be a phenomenon analogous to visual hallucinations in patients with dementia with Lewy bodies and in Parkinson’s disease without dementia ⁵. Face pareidolia seems to be a universal phenomenon, and it has been observed even in autistic individuals who may show a deficit in face processing ⁶. Akechi et al. ⁷ recorded the N170 event-related potential in age-matched autism spectrum disorder (ASD) and control adolescents during the perception of objects and faces-in-objects. The results showed that both the autism spectrum disorder (ASD) adolescents and the healthy developing adolescents showed highly similar face-likeness ratings. Both groups also showed enhanced face-sensitive N170 amplitudes to face-like objects vs. objects. The authors concluded that both autism spectrum disorder (ASD) adolescents and the healthy developing adolescents individuals exhibit perceptual and neural sensitivity to face-like features in objects ⁷.

Compared with other types of illusion, pareidolia is unique in how the illusion often becomes more intense with increased attention to it. Similar neural processes trigger pareidolic illusions and visual hallucinations, which has led to speculation that pareidolia represents a susceptibility to visual hallucinations ⁵.

What causes pareidolia

The underlying cause of pareidolia is unknown. Comprehensive studies ⁸ have investigated which brain regions participate in the processing of real-face and face-pareidolia stimuli. However, the brain regions that exhibit activation during these processes have yet to be fully determined. Despite their lack of experience, newborn infants preferentially look at faces over other objects, which suggests that the brain regions responsible for face recognition are functioning at birth ⁹. However, other studies have suggested that face-specific regions become functional following experience with visual stimuli during one’s lifetime ¹⁰. Nonetheless, faces remain the most important stimuli for visual perception studies. Furthermore, face pareidolia is more common than object pareidolia because the human visual system is highly sensitive to and adapted for the recognition of faces.

Kanwisher et al. ¹¹ found that face stimuli evoked activation on the lateral side of the mid-fusiform gyrus, also called the Fusiform Face Area, and a Functional Magnetic Resonance Imaging (fMRI) study by Kanwisher and Yovel ¹² showed that the fusiform regions exhibited greater activation to faces than to letter strings or textures. Recent studies investigating the Fusiform Face Area have shown that this region exhibits significant activation for real faces as well as during face pareidolia ¹³. Face pareidolia refers to the attribution of real face traits to non-face objects due to illusory perceptions. This phenomenon typically occurs when non-face stimuli erroneously activate a connection between visual input areas and internal representations ¹⁴.

Liu et al. ¹⁵ investigated the neural and behavioural correlates of face pareidolia and showed that the Fusiform Face Area played a crucial role in the perception of real faces as well as the processing of illusory face perceptions. These authors also reported that the Fusiform Face Area showed higher activation during face pareidolia than during letter pareidolia. The neural mechanisms underlying face-pareidolia perception were proposed to include an interaction between bottom-up processing and top-down processing ¹⁵ and accordingly, the brain analyses in this study revealed that both frontal and occipitotemporal regions were important for processing face pareidolia. Occipitotemporal regions, including the primary visual cortex, enable the perception of visual input and are highly associated with bottom-up processing, whereas the frontal cortex is responsible for reasoning and is highly associated with top-down processing. When presented with illusory face stimuli that have components similar to normal faces, such as eyes and mouths, humans engage in top-down processing that creates associations with previous knowledge about real faces, which results in interpreting an illusory face as a real face.

Pareidolia test

The pareidolia test is a tool that evokes visual hallucination-like illusions, and these illusions may be a surrogate marker of visual hallucinations in dementia with Lewy bodies ². Dementia with Lewy bodies is the second most common form of neurodegenerative dementia, and it accounts for ~9.7% of people with dementia in population-based studies and ~24.7% of people with dementia in clinic-based studies ¹⁶. Patients with Lewy bodies dementia with suffer more severe functional impairments and a higher mortality rate than patients with Alzheimer’s disease ¹⁷. Caregiver burden and distress are also greater in dementia with Lewy bodies than Alzheimer’s disease, which leads to earlier institutionalization of patients with dementia with Lewy bodies in nursing homes ¹⁸. Therefore, patients with dementia with Lewy bodies are associated with higher healthcare costs than those with Alzheimer’s disease ¹⁹. Early diagnosis and intervention are required to improve the quality of life of patients and caregivers and to reduce the economic burden to society, but this need is currently unmet. Previous clinico-pathological studies demonstrated that the specificity of the clinical diagnostic criteria for dementia with Lewy bodies is consistently high (over 90%), but its sensitivity is extremely low (12.1–83.0%) ²⁰, suggesting that underdiagnosis is a significant issue.

Visual hallucinations are one of the three core symptoms in the diagnostic criteria for dementia with Lewy bodies ²¹, and are observed in over 70% of clinically diagnosed dementia with Lewy bodies patients ²². Visual hallucinations are defined as false perception that arise independently from actual visual scenes, whereas visual illusions are misperceptions that result from distortion of actual visual scenes. However, the distinction between visual hallucinations and visual illusions are often ambiguous because patients see things whenever they are awake and have their eyes open.

There are two versions of the pareidolia test, the scene and noise pareidolia tests (https://ndownloader.figshare.com/files/4977159) ²³. Although both test versions evoked phenomenologically similar pareidolic illusions, most of which were images of humans and animals, they exhibited distinct noso- and psychometric profiles. The scene pareidolia test demonstrated an excellent ability to discriminate dementia with Lewy bodies from Alzheimer’s disease, but it was poorly correlated with clinical visual hallucinations ²⁴. The noise pareidolia test correlated well with visual hallucinations, but its utility for differentiating between dementia with Lewy bodies and Alzheimer’s disease was inferior to the scene pareidolia test ²³. The scene and noise pareidolia tests may reflect different aspects of psychological or neurological mechanisms of pareidolias. We combined both tests in the current study to take advantage of the assets of each test version. We also abbreviated the test and simplified its administration and scoring procedures to improve its usefulness in clinical settings. The pareidolia score, a composite score of the scene and noise pareidolia tests, exhibited an excellent inter-rater/test-retest reliability, good correlation with clinical hallucinations and a better balance of sensitivity and specificity compared with each of the two test versions alone. The mean administration time was approximately 15 minutes, which is acceptable in clinical settings. Therefore, the current version of the pareidolia test may be a helpful tool in dementia clinics.

The pareidolia score and noise pareidolia test significantly correlated with the Neuropsychiatric Inventory hallucinations score but were not correlated with measures of global cognitive ability, such as the Mini-mental State Examination and Addenbrooke’s Cognitive Examination-Revised total scores. The pareidolia score and noise pareidolia test may serve as a reliable surrogate marker of visual hallucinations in dementia patients with mild to moderate cognitive impairment, which was the level of cognitive severity of subjects in the current study. The pareidolia score exhibited better sensitivity to discriminate dementia with Lewy bodies from Alzheimer’s disease compared with the noise pareidolia test alone. However, the use of the noise pareidolia test independently may be a good option to measure treatment responses because the noise pareidolia test exhibited a good correlation with clinical visual hallucinations. By contrast, we found no significant correlation between the scene pareidolia test and Neuropsychiatric Inventory hallucinations score, which is consistent with a previous study ²⁴. This result likely occurred because not only dementia with Lewy bodies patients with frank hallucinations but also those without visual hallucinations produced illusory responses in this test. The reduced number of test stimuli and simplification of the scoring procedure may also have reduced the quantitativeness of the test at the expense of the ease of test administration. A subset of Alzheimer’s disease patients produced as many illusory responses as dementia with Lewy bodies patients. The low sensitivity of the current clinical diagnostic criteria for dementia with Lewy bodies suggest that this result does not merely represent false positives, but it may predict the future development of full-blown dementia with Lewy bodies symptoms ²⁰. The prediction ability of the scene pareidolia test for conversion to clinical dementia with Lewy bodies should be addressed in future studies.

The inter-rater/test-retest reliabilities of the pareidolia score and noise pareidolia test were excellent, but the reliabilities of the scene pareidolia test remained moderate. We conducted two supplementary analyses to identify the factors that were associated with inter-rater/test-retest variability. In the first analysis, intra-class correlation coefficient for the scene pareidolia test, noise pareidolia test and pareidolia score were calculated separately in 15 dementia with Lewy bodies patients and 15 Alzheimer’s disease patients. In the second analysis, we presented the identical video-recorded patient responses (5 dementia with Lewy bodies and 5 Alzheimer’s disease) in the scene pareidolia test to two examiners and compared their scores. Intra-class correlation coefficient was significantly lower in dementia with Lewy bodies patients than Alzheimer’s disease patients in the scene pareidolia test and the inter-rater agreement on the scene pareidolia test was excellent (0.99). These findings suggest that cognitive fluctuations in dementia with Lewy bodies patients and the manner of test administration, including instructions and feedback, may be the critical factors of inter-rater/test-retest variability in the scene pareidolia test.

The setting of cut-off scores for quantitative cognitive tests is helpful from a differential diagnosis perspective. The receiver operating characteristic analyses indicate that the optimal cut-off for the pareidolia score is 4/5 for the differentiation of dementia with Lewy bodies from Alzheimer’s disease and 2/3 for the differentiation of dementia with Lewy bodies and healthy controls. However, these cut-off scores were calculated based on a small number of patients with limited forms of dementing disorders and with a narrow range of severity. Uncritical applications of these cut-off scores to individual patients may lead to an erroneous diagnosis and inappropriate treatment. Scores that are just below or above the cut-off scores should be interpreted with consideration of other clinical information. For example, if a multidomain non-amnestic mild cognitive impairment patient who exhibits prominent cognitive fluctuations scores 2 on the pareidolia score, a false negative may be considered. By contrast, a clinician may consider the possibility of a false positive when a severely amnesic dementia patient scores 6 without any other dementia with Lewy bodies-like clinical data. The validity of the cut-off scores proposed here, and the utility of the pareidolia test in a broader sense, should be examined in studies with a larger number of patients and in more diverse patient population including other types of dementias and psychiatric illnesses.

1. Akdeniz G, Toker S, Atli I. Neural mechanisms underlying visual pareidolia processing: An fMRI study. Pak J Med Sci. 2018;34(6):1560-1566. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6290235/[↩]
2. Mamiya Y, Nishio Y, Watanabe H, et al. The Pareidolia Test: A Simple Neuropsychological Test Measuring Visual Hallucination-Like Illusions. PLoS One. 2016;11(5):e0154713. Published 2016 May 12. doi:10.1371/journal.pone.0154713 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4865118/[↩][↩]
3. Uchiyama M, Nishio Y, Yokoi K, et al. Pareidolia in Parkinson’s disease without dementia: a positron emission tomography study. Parkinsonism Relat Disord 2015;21:603–9[↩]
4. Kato M, Mugitani R. Pareidolia in infants. PLoS One. 2015;10(2):e0118539. Published 2015 Feb 17. doi:10.1371/journal.pone.0118539 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4331561/[↩]
5. Uchiyama M, Nishio Y, Yokoi K, et al. Pareidolia in Parkinson’s disease without dementia: a positron emission tomography study. Parkinsonism Relat Disord 2015;21:603–9. https://www.ncbi.nlm.nih.gov/pubmed/25864093[↩][↩]
6. Proverbio AM, Galli J. Women are better at seeing faces where there are none: an ERP study of face pareidolia. Soc Cogn Affect Neurosci. 2016;11(9):1501-12. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5015812[↩]
7. Neural and behavioural responses to face-likeness of objects in adolescents with autism spectrum disorder. Akechi H, Kikuchi Y, Tojo Y, Osanai H, Hasegawa T. Sci Rep. 2014 Jan 27; 4():3874.[↩][↩]
8. Taubert J, Wardle SG, Flessert M, Leopold DA, Ungerleider LG. Face pareidolia in the rhesus monkey. Curr Biol. 2017;27(16):2505–2509. doi.org/10.1016/j.cub.2017.06.075[↩]
9. Cassia VM, Turati C, Simion F. Can a nonspecific bias toward top-heavy patterns explain newborns face preference? Psychol Sci. 2004;15(6):379–383. doi:10.1111/j.0956-7976.2004.00688.x[↩]
10. Bukach CM, Gauthier I, Tarr MJ. Beyond faces and modularity:the power of an expertise framework. Trends Cogn Sci. 2006;10(4):159–166. doi:10.1016/j.tics.2006.02.004[↩]
11. Kanwisher N, McDermott J, Chun MM. The fusiform face area:a module in human extra striate cortex specialized for face perception. J Neurosci. 1997;17:4302–4311. doi.org/10.1523/JNEUROSCI[↩]
12. Kanwisher N, Yovel G. The fusiform face area:a cortical region specialized for the perception of faces. Philos Trans R Soc Lond B Biol Sci. 2006;361:2109–2128. doi:10.1098/rstb.2006.1934[↩]
13. Proverbio AM, Galli J. Women are better at seeing faces where there are none:an ERP study of face pareidolia. Soc Cogn Affect Neurosci. 2016;11(9):1501–1502. doi:10.1093/scan/nsw064[↩]
14. Collerton D, Perry E, McKeith I. Why people see things that are not there:a novel perception and attention deficit model for recurrent complex visual hallucinations. Behav Brain Sci. 2005;28:737–794. doi.org/10.1017/S0140525X05000130[↩]
15. Liu J, Li J, Feng L, Li L, Tian J, Lee K. Seeing Jesus in a toast:neural and behavioural correlates of face pareidolia. Cortex. 2014;53:60–77. doi:10.1016/j.cortex[↩][↩]
16. Vann Jones SA, O’Brien JT (2014) The prevalence and incidence of dementia with Lewy bodies: a systematic review of population and clinical studies. Psychol Med 44: 673–683. 10.1017/S0033291713000494[↩]
17. Ricci M, Guidoni SV, Sepe-Monti M, Bomboi G, Antonini G, et al. (2009) Clinical findings, functional abilities and caregiver distress in the early stage of dementia with Lewy bodies (DLB) and Alzheimer’s disease (AD). Arch Gerontol Geriatr 49: e101–104. 10.1016/j.archger.2008.10.001[↩]
18. Rongve A, Vossius C, Nore S, Testad I, Aarsland D (2014) Time until nursing home admission in people with mild dementia: comparison of dementia with Lewy bodies and Alzheimer’s dementia. Int J Geriatr Psychiatry 29: 392–398. 10.1002/gps.4015[↩]
19. Bostrom F, Jonsson L, Minthon L, Londos E (2007) Patients with Lewy body dementia use more resources than those with Alzheimer’s disease. Int J Geriatr Psychiatry 22: 713–719.[↩]
20. Nelson PT, Jicha GA, Kryscio RJ, Abner EL, Schmitt FA, et al. (2010) Low sensitivity in clinical diagnoses of dementia with Lewy bodies. J Neurol 257: 359–366. 10.1007/s00415-009-5324-y[↩][↩]
21. McKeith IG, Dickson DW, Lowe J, Emre M, O’Brien JT, et al. (2005) Diagnosis and management of dementia with Lewy bodies: third report of the DLB Consortium. Neurology 65: 1863–1872[↩]
22. Nagahama Y, Okina T, Suzuki N, Matsuda M, Fukao K, et al. (2007) Classification of psychotic symptoms in dementia with Lewy bodies. Am J Geriatr Psychiatry 15: 961–967.[↩]
23. Yokoi K, Nishio Y, Uchiyama M, Shimomura T, Iizuka O, et al. (2014) Hallucinators find meaning in noises: pareidolic illusions in dementia with Lewy bodies. Neuropsychologia 56: 245–254. 10.1016/j.neuropsychologia.2014.01.017[↩][↩]
24. Uchiyama M, Nishio Y, Yokoi K, Hirayama K, Imamura T, et al. (2012) Pareidolias: complex visual illusions in dementia with Lewy bodies. Brain 135: 2458–2469. 10.1093/brain/aws126[↩][↩]