Key Points
Question
To what degree are societal members able to identify paralysis of varying severity and localize the defect on the face?
Findings
In this prospective observational study of 380 casual observers, facial paralysis was identified in 249 (34.6%) of 719 facial photographs with low-grade paralysis, 448 (63.2%) of 709 with medium-grade paralysis, and 696 (96.7%) of 720 with high-grade paralysis, and participants tended to identify facial paralysis more accurately in smiling vs repose faces.
Meaning
The ability of individuals to identify the presence of facial paralysis increased as paralysis severity increased and smiling increased accurate identification; however, even when individuals could identify paralysis, they were not necessarily able to accurately localize the paralysis on a face.
This observational study examines the ability of casual observers to identify paralysis of varying degrees of severity and to localize the defect on the face.
Abstract
Importance
When able to identify facial paralysis, members of society regard individuals with facial paralysis differently. They perceive a decrease in attractiveness, more negative affect, and lower quality of life. However, the ability of lay people in society to accurately identify the presence of facial paralysis has not yet been defined.
Objective
To determine societal members’ ability to (1) identify paralysis in varying degrees of paralysis severity and (2) localize the defect on the face.
Design, Setting, and Participants
A prospective observational study conducted in an academic tertiary referral center using a group of 380 casual observers was carried out.
Main Outcomes and Measures
Surveys were designed containing smiling and repose images of normal faces and faces with unilateral facial paralysis of 3 severity levels (low, medium, and high) as categorized by House-Brackmann (HB) grade. The photographs were then shown to casual observers in a web-based survey. After reviewing both normal faces and faces with varying degrees of paralysis, they then indicated (1) whether paralysis was present and (2) if so, where the paralysis was on the face.
Results
A total of 380 participants (267 [70.3%] women and 113 [29.7%] men with a mean [SD] age of 29 [12] years) successfully completed the survey, viewing 2860 facial photographs in aggregate. The accuracy rate of identifying paralysis increased from low-grade through high-grade paralysis. Facial paralysis was identified in 249 (34.6%) of 719 facial photographs with low-grade paralysis, 448 (63.2%) of 709 with medium-grade paralysis, and 696 (96.7%) of 720 with high-grade paralysis (χ2 = 912.6, P < .001); 6.2% (44/731) of normal faces were incorrectly identified as having paralysis (χ2 = 912.6, P < .001). Participants correctly localized paralysis in 157 (63.0%) of 249 low-grade photographs, 307 (68.5%) of 448 medium-grade photographs, and 554 (79.6%) of 696 high-grade photographs (χ2 = 32.5, P < .001). In general, participants tended to identify facial paralysis more accurately in smiling vs repose faces (48.6% vs 20.6%, 92.4% vs 33.7%, and 96.7% vs 96.6% in low-, medium-, and high-grade paralysis, respectively) (χ2 = 62.2, P < .001; χ2 = 262.6, P < .001; χ2 = 0.0, P = .96, respectively).
Conclusions and Relevance
The ability of individuals to identify the presence of facial paralysis increased as paralysis severity increased. Further, smiling increased accurate identification. However, even when individuals can identify paralysis, they are not necessarily able to accurately localize the paralysis on a face. This may speak to a phenomenon in which perception of a facial defect comes from a holistic interpretation of a face, rather than a clinically accurate specification of the defect location. These findings are important in the future counseling of patients.
Level of Evidence
NA.
Introduction
Patients with facial paralysis often have diminished quality of life owing to functional deficits of facial expression because they are perceived to have more negative affect and decreased attractiveness levels.1,2,3,4,5 Facial paralysis compromises a key communicative function of human emotion, which is to relay information between individuals and to adapt one’s actions accordingly.6 It is a unique condition in which self-perceived quality of life is inextricably linked to how other members of society perceive oneself,2 yet the perspective of the societal member is a difficult one to measure.7 It has been shown that when able to detect paralysis and imagine the paralysis as inflicting their own faces, members of society are willing to pay high levels of money to reanimate their face.4 The willingness to pay for correction as well as societally perceived quality of life fluctuates depending on the severity of paralysis as defined by House-Brackmann (HB) grade and on the facial expression perceived (eg, smiling or in repose).
Although it has been shown that observers perceive facial paralysis to be an important condition to repair with significant health utility implications, the rate at which casual observers can detect paralysis has not yet been quantified. In fact, a previous study has shown that some casual observers may even mistake normal faces as having paralysis, and that there is an inflection point of paralysis severity after which one’s willingness to pay increases dramatically.4 This posits that a layperson’s ability to identify paralysis is unreliable below a certain threshold, and that some level of facial asymmetry must exist for a casual observer to appropriately interpret a face as having a defect.8 In addition, this is an important investigative question considering that clinical counseling for repair of facial paralysis and discussion of the patient’s goals often center around what defects surgeons and patients feel can be detected by other members of society. One cannot assume that paralysis detectable by a facial plastic surgeon can also be detected by the societal member, and the field of facial plastic surgery is moving toward a more objective understanding of facial deformities.9
The aims of this study are to investigate how accurately casual observers can (1) identify paralysis in varying degrees of paralysis severity and (2) to localize the defect on the face.
Methods
Participants
The Johns Hopkins Medicine institutional review board approved this study. Written informed consent was obtained and participants were not compensated. Surveys posted on public access websites were used to recruit casual observers as participants. Exclusion criteria included participants younger than 18 years and those with schizophrenia or autism spectrum disorders owing to differences in ways these individuals perceive faces.10,11 Forty participants were excluded, and complete surveys were collected from 380 participants. Participants were recruited from July 20, 2015, to September 26, 2015. Demographic profiles of participants are presented in Table 1.
Table 1. Demographics of Study Participants.
| Characteristic | No. (%) |
|---|---|
| Age, mean (SD), y | 29 (12) |
| Sex | |
| Female | 267 (70.3) |
| Male | 113 (29.7) |
| Race | |
| Asian | 111 (29.2) |
| African American | 31 (8.2) |
| White | 210 (55.3) |
| American Indian | 7 (1.8) |
| Pacific Islander | 4 (1.0) |
| Other | 17 (4.5) |
| Hispanic/Latino | 22 (5.8) |
| Education | |
| Some high school | 2 (0.5) |
| High school/GED | 25 (6.6) |
| Some college | 52 (13.7) |
| 2-year college degree | 11 (2.9) |
| 4-year college degree | 198 (52.1) |
| Master’s degree | 75 (19.7) |
| Doctoral degree | 17 (4.5) |
| Annual household income, $K USD | |
| <25 | 117 (30.8) |
| 25-49 | 74 (19.5) |
| 50-74 | 48 (12.6) |
| 75-99 | 38 (10.0) |
| 100-149 | 37 (9.7) |
| 150-199 | 25 (6.6) |
| ≥200 | 29 (7.6) |
| NA | 12 (3.2) |
| Facial paralysis | |
| Personal facial paralysis | 8 (2.1) |
| Relatives with facial paralysis | 24 (6.3) |
Survey Instrument and Data Analysis
Photographs of 16 individuals (4 individuals without paralysis serving as controls and 12 individuals with unilateral facial paralysis) were queried from the Johns Hopkins facial plastic surgery image archive. Of the 12 individuals with facial paralysis, there were 4 individuals in each of 3 categories of facial paralysis: low, medium, and high. Low-grade paralysis was categorized as HB grade 2, medium-grade as HB grades 3 to 4, and high-grade as HB grades 5 to 6. Four unique surveys were then generated with 8 photographs each. The composition of each survey included 2 photographs in each of 4 categories: normal, low-grade, medium-grade, and high-grade paralysis. We included 1 smiling and 1 repose photograph for each of the 4 categories. Surveys were generated using Qualtrics Online Survey Software (Qualtrics LLC).12 During the design of the survey, a region was drawn over the side of the face with paralysis if a patient was determined to be HB grade 2 and above. These regions captured portions of the face in which a trained facial plastic surgeon could identify paralysis (eg, regions of mouth, nasolabial fold, eyes). The software remembers the x and y coordinates of the regions that were drawn but does not display these drawn areas to participants. Examples of faces shown to participants and regions drawn with software are shown in the Figure.
Figure. Survey Photograph of Facial Paralysis.
A, Example photo of patient with medium-grade facial paralysis with smiling facial expression shown to participants. B, Photo input into Qualtrics Survey Software with a region drawn to capture areas identified as exhibiting facial paralysis. This drawn region is not shown to casual observers. If a casual observer clicks within the predesignated area, the participant is marked as having accurately identified the location of paralysis.
Each individual was only shown once within a survey, and no photograph was duplicated among the 4 surveys. Participants were asked whether they saw paralysis in each face. If they answered “yes,” they were then prompted to click on the portion of the face in which they saw the paralysis. The Qualtrics Software recorded the x and y coordinates of the click and generated a binary “yes” or “no” in data collection that indicated to the data analyst whether the participant clicked within the predrawn region containing paralysis. Data were analyzed using Stata statistical software (version 13 SE, Stata Corp).
Results
A total of 380 completed surveys were analyzed, amounting to a total of 2860 facial photographs viewed. The sex of participants was skewed toward female (267 [70.26%] participants), and the mean age was 29 years. Income distribution was representative of the US population.13
The aim of this study was to understand how accurate the layperson was at identifying (1) whether a face had paralysis and (2) the location of the paralysis. This study was also interested in elucidating how 2 specific characteristics of a patient affect the casual observer’s ability to accomplish these tasks: (A) severity of facial paralysis (as defined by the HB scale) and (B) facial positioning (defined as a smiling face or a face in repose).
Participants correctly identified normal faces as not having paralysis 93.8% of the time. The ability of participants to identify facial paralysis increases as the paralysis severity increases. Facial paralysis was identified in 249 (34.6%) of 719 facial photographs with low-grade paralysis, 448 (63.2%) of 709 with medium-grade paralysis, and 696 (96.7%) of 720 with high-grade paralysis (χ2 = 912.6, P < .001). In addition, participants’ ability to identify paralysis increased if the facial photograph was one of a smiling face vs a face in repose. Among low-grade paralysis photos, participants accurately identified 175 (48.6%) of 360 smiling photographs as having paralysis vs 74 (20.6%) of 359 repose photographs as having paralysis (χ2 = 62.2, P < .001). Among medium-grade paralysis, participants accurately identified paralysis in 329 (92.4%) of 356 smiling photographs vs 119 (33.7%) of 353 repose photographs (χ2 = 262.6, P < .001). In the high-grade paralysis category, participants accurately identified paralysis in 352 (96.7%) of 364 smiling photographs vs 344 (96.6%) of 356 repose photographs, although this was not found to be statistically significant (χ2 = 0.0, P = .96). Complete results are shown in Table 2.
Table 2. Casual Observer Identification of Presence of Facial Paralysis.
| Variable | Faces Correctly Identified as Normal, No. (%) | Faces Incorrectly Identified as Paralyzed, No. (%) | χ2 | P Value |
|---|---|---|---|---|
| Normal face | 669 (93.8) | 44 (6.2) | 912.6 | <.001 |
| Low-grade paralysisa | 249 (34.6) | 470 (65.4) | 912.6 | <.001 |
| Smiling | 175 (48.6) | 185 (51.4) | 62.2 | <.001 |
| In repose | 74 (20.6) | 285 (79.4) | ||
| Medium-grade paralysisb | 448 (63.2) | 261 (36.8) | 912.6 | <.001 |
| Smiling | 329 (92.4) | 27 (7.6) | 262.6 | <.001 |
| In repose | 119 (33.7) | 234 (66.3) | ||
| High-grade paralysisc | 696 (96.7) | 24 (3.3) | 912.6 | <.001 |
| Smiling | 352 (96.7) | 12 (3.3) | 0.0 | .96 |
| In repose | 344 (96.6) | 12 (3.4) |
Low-grade paralysis is defined as House-Brackmann (HB) grade 2.
Medium-grade paralysis is defined as HB grades 3 and 4.
High-grade paralysis is defined as HB grades 5 and 6.
To understand whether participants were able to correctly localize the area of paralysis after indicating that they saw paralysis, the data were analyzed to see whether participants correctly clicked predrawn regions of paralysis in the photographs. For all grades of paralysis, most participants who correctly identified facial paralysis also correctly identified the location of the facial paralysis. The accuracy of localization also increased with increasing grade of paralysis. Participants correctly localized paralysis in 157 (63.0%) of 249 low-grade photographs, 307 (68.5%) of 448 medium-grade photographs, and 554 (79.6%) of 696 high-grade photographs (χ2 = 32.5, P < .001). Complete results are shown in Table 3.
Table 3. Casual Observer Identification of Facial Paralysis Location.
| Grade | Faces With FP Correctly Localized, No. (%) | Faces With FP Incorrectly Localized, No. (%) | χ2 | P Value |
|---|---|---|---|---|
| Low-grade paralysisa | 157 (63.0) | 92 (36.9) | 32.5 | <.001 |
| Medium-grade paralysisb | 307 (68.5) | 141 (31.5) | ||
| High-grade paralysisc | 554 (79.6) | 142 (20.4) |
Low-grade paralysis is defined as House-Brackmann (HB) grade 2.
Medium-grade paralysis is defined as HB grades 3 and 4.
High-grade paralysis is defined as HB grades 5 and 6.
Discussion
This is the first study, to our knowledge, conducted to quantify the accuracy of the casual observer in identifying both the presence of facial paralysis in a patient graded by severity and facial positioning, as well as the accuracy of the observer in identifying the location of paralysis.
The results support our initial hypothesis that societal members are more likely to identify paralysis if the paralysis is severe. Within each paralysis grade, societal members were also more likely to identify paralysis in a smiling face vs a face in repose. This is consistent with past quality of life studies in which societal members deem perceived quality of life to be significantly worse in high-grade vs medium- and low-grade paralyses; casual observers rated a more negative affect, decreased attractiveness, and decreased health utility using 3 separate quantification methods with increasing grades of paralysis.4 Su et al4 had previously shown that a societal member’s willingness-to-pay for facial reanimation increased nearly exponentially from low- to high-grade paralysis, raising the question of resource allocation to concentrate on higher-grade vs lower-grade paralysis. In a similar trend, this study showed that only a minority of photographs with low-grade paralysis were identified as having paralysis (34.63%) whereas most medium-grade and high-grade paralysis photographs were accurately identified. This again reinforces the need for counseling to direct resources and surgical interventions toward medium- and high-grade paralyses while advising patients of low-grade paralysis that their lesions may be undetectable to the casual observer.
With regard to identification of location of paralysis, our initial hypotheses were first that the casual observer would not be as accurate in identifying location as facial plastic surgery experts, and second that unilateral facial paralysis can have an overall effect on a face not limited to the side of paralysis, which can lead to further inaccuracy in location identification. Results were consistent with this hypothesis. Notably, among the photographs correctly identified as having paralysis, paralysis was inaccurately localized in 92 (37.0%) of 249 low-grade paralysis photographs and in 141 (31.5%) of 448 medium-grade paralysis photographs. This may be owing to the fact that the paralyzed side has weaker movements; thus, observers may sometimes perceive that the side with normal motion is the one with the defective lesion, because it is that side that is actively distorting the face into an expression. Walker-Smith et al14 published a seminal paper on scanpaths—that is, the characteristic path that one’s gaze follows when regarding another person’s face. Their study found that eye movements follow a highly conserved and regular sequence to aid in both facial recognition and visual memory.14 This path was confined to the eyes-nose-mouth area on bilateral sides of the face, resulting in the memory of 1 singular complete face. Such a conserved path of foveation will pick up a defect, no matter on which side, that delivers a singular perception of a face—eg, “normal” or “paralyzed.” This study suggests that one need not accurately define the origin of asymmetry to scan over a lesion and identify that a defect exists. From low- to high-grade paralysis, accuracy in location identification similarly increased with increasing severity of paralysis. The clinical implication of this is for counseling purposes—that facial reanimation and other corrective therapies are not necessarily to repair an obvious lesion to a casual observer, but rather to improve the overall perception of a face.
The findings of this study emphasize that the perception of a paralyzed face from the prospective of a casual observer is unlike that of a facial plastic expert, and lower grades of paralysis may go undetected. Importantly, it also showed that the ability to localize the paralysis does not preclude a casual observer from recognizing that a lesion exists affecting the overall face, implying that a unilateral lesion can negatively affect areas of a face unaffected by paralysis.
Limitations
This study grossly analyzes how accurate a casual observer is in locating regions of paralysis. The regions drawn with the Qualtrics survey tool were large and sometimes involved an entire hemifacial area. It would be valuable to explore the coordinates that observers clicked on more closely to see whether there is a pattern of observers identifying specific points of a face as being abnormal (ie, nasolabial fold, lagophthalmos, incomplete smile). Though Walker-Smith et al14 has delineated the typical eye tracking movement on perception of a face, it would be helpful to quantify which part of the face is the most prominent abnormality identified by an observer.
Another limitation is the use of photos in this study, rather than the incorporation of dynamic videos. This decision was made to decrease survey fatigue and time required to complete the survey; in addition, the design of the survey was to account for the fact that the perceived face of a stranger is usually in repose for more duration of time than in smile or motion. However, it should be noted that ability to detect both presence and location of paralysis would likely increase if casual observers were able to see dynamic facial motion.
Finally, the participant population was skewed toward younger individuals and the female sex.
Conclusions
The ability of casual observers to identify the presence of facial paralysis increases as paralysis severity increases and in smiling faces vs faces in repose. However, even when individuals can identify paralysis, they are not necessarily able to accurately localize the paralysis on a face. This may speak to a phenomenon in which perception of a facial defect comes from a holistic interpretation of a face, rather than a clinically accurate specification of the defect location.
References
- 1.Kleiss IJ, Hohman MH, Susarla SM, Marres HA, Hadlock TA. Health-related quality of life in 794 patients with a peripheral facial palsy using the FaCE Scale: a retrospective cohort study. Clin Otolaryngol. 2015;40(6):651-656. [DOI] [PubMed] [Google Scholar]
- 2.Ishii L, Godoy A, Encarnacion CO, Byrne PJ, Boahene KD, Ishii M. Not just another face in the crowd: society’s perceptions of facial paralysis. Laryngoscope. 2012;122(3):533-538. [DOI] [PubMed] [Google Scholar]
- 3.Coulson SE, O’Dwyer NJ, Adams RD, Croxson GR. Expression of emotion and quality of life after facial nerve paralysis. Otology & neurotology. 2004;25(6):1014-1019. [DOI] [PubMed] [Google Scholar]
- 4.Su P, Ishii LE, Joseph A, et al. Societal value of surgery for facial reanimation. JAMA Facial Plast Surg. 2017;19(2):139-146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ryzenman JM, Pensak ML, Tew JM Jr. Facial paralysis and surgical rehabilitation: a quality of life analysis in a cohort of 1,595 patients after acoustic neuroma surgery. Otology & neurotology. 2005;26(3):516-521. [DOI] [PubMed] [Google Scholar]
- 6.Walker-Andrews AS. Infants’ perception of expressive behaviors: differentiation of multimodal information. Psychol Bull. 1997;121(3):437-456. [DOI] [PubMed] [Google Scholar]
- 7.Hadlock T. Standard Outcome Measures in Facial Paralysis: Getting on the Same Page. JAMA Facial Plast Surg. 2016;18(2):85-86. [DOI] [PubMed] [Google Scholar]
- 8.Chu EA, Farrag TY, Ishii LE, Byrne PJ. Threshold of visual perception of facial asymmetry in a facial paralysis model. Arch Facial Plast Surg. 2011;13(1):14-19. [DOI] [PubMed] [Google Scholar]
- 9.Ishii LE. Moving toward objective measurement of facial deformities: exploring a third domain of social perception. JAMA Facial Plast Surg. 2015;17(3):189-190. [DOI] [PubMed] [Google Scholar]
- 10.Nakano T, Tanaka K, Endo Y, et al. Atypical gaze patterns in children and adults with autism spectrum disorders dissociated from developmental changes in gaze behaviour. Proc Biol Sci. 2010;277(1696):2935-2943. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Manor BR, Gordon E, Williams LM, et al. Eye movements reflect impaired face processing in patients with schizophrenia. Biol Psychiatry. 1999;46(7):963-969. [DOI] [PubMed] [Google Scholar]
- 12.Qualtrics [computer program]. Provo, Utah, USA2015.
- 13.Carmen DeNavas-Walt BDP. Income and Poverty in the United States: 2014. United States Census Bureau; 2015. [Google Scholar]
- 14.Walker-Smith GJGA, Gale AG, Findlay JM. Eye movement strategies involved in face perception. Perception. 1977;6(3):313-326. [DOI] [PubMed] [Google Scholar]
