1. Selection of participants - A total of 50 healthy adult participants were included in this study. A power calculation using G*power was done to estimate the sample size for our study. Based on our primary research question, we planned to do a MANOVA with three response variables (averaged semantic memory scores, averaged free recollection and general probing based memory scores, and averaged specific probing based memory scores). We had 4 conditions based on the factors of group (child face filter, no face filter) and visuo-motor synchrony (synchronous, asynchronous). Using these parameters, we aimed to have a medium effect size (f2 = .15), a error probability of .05, and a power of .90. Following the power calculation, we estimated our total sample size to be 48, in response to which we collected data from 50 participants. The mean age of the sample was 28.14 (SD = 4.94) years and there were 34 female participants and 16 males. None of the participants had any self-declared history of previous psychiatric or neurological disorder. The participants were only included if they had self-declared access to Wi-Fi with over 25 Mbps download speed, access to video conferencing software Zoom, had a laptop or personal computer with screen size more than 13 inches and with a webcam. Online informed consent was acquired by all the participants prior to participation. Ethical approval was granted by the Science and Engineering Faculty Research Ethics Panel (FREP) of Anglia Ruskin University.
2. Experimental design and overview of the procedure - The online experiment was created using the software Gorilla experiment builder 1. The researcher was in contact with the participant during the experiment using the video conferencing software Zoom 2. Every participant participated in four experimental conditions (table 1) based on the combinations of visuo-motor synchrony (synchronous, asynchronous) during the enfacement illusion and type of memory probed (recent, childhood). Participants were provided with a link that opened the Gorilla platform on their browser 1. On the platform they read through the information sheet and provided informed consent before starting the experiment. They were then directed to a zoom call in which the researcher was already present. Following this, they were instructed to make the practice head movements for 90 seconds. They were then randomly assigned to the child face group or no filter group. This randomisation was counterbalanced to have 25 participants assigned to the child face group and the remaining 25 to the no filter group. Each participant then took part in the four experimental conditions in a randomised and counterbalanced sequence. In each condition, the participants were instructed to make the head movement (synchronous or asynchronous) for 90 seconds while watching the video feed of their own face. Immediately after the head movements, the participants completed the enfacement illusion questionnaire. Immediately after that the AMI was conducted, and the interview was recorded for it to be later scored and analysed. After the completion of the four experimental conditions, the participants were debriefed, and the experiment concluded.
3. Group allocation - The participants were randomly allocated into two groups (child face filter, no face filter). The groups were found to be appropriately matched for age and gender. The mean age for child face filter group was 28.28 (SD = 6.161) years and the mean age for no face filter group was 28.00 (SD = 3.571) years. There was non-normal distribution of age for child face filter group (W = .819, p < .001) so we conducted a Mann-Whitney U test to assess any differences in age between the groups. We found no significant difference in age for the two groups (U = 286.00, p = .605). We performed a chi squared test to assess the matching of gender for the two groups and found that the males and females were in the same proportions for the two groups (X2(1, N = 50) = .00, p = 1.000) as shown in fig. 1.
4. Implementation of enfacement illusion - In the ‘Child face filter’ group (n=25), participants viewed a ‘younger version’ of their own face as a live video. This was created by using ‘baby face’ filter of the Snap Camera software as shown in fig. 2. The other group (‘no face filter’ group; n=25) viewed a live video of their face on the screen without any morphing or manipulation. The enfacement illusion was created by manipulating the visuo-motor synchrony between the physical head movements of the participants and the video feed that the participants saw while making the head movements. In a practise phase before the main experiment, the participants were first shown a pre-recorded video of an actor moving their head with the repetitive motion of left-centre-right-centre around the vertical central axis at 45 bpm (beats per minute) for 90 seconds. The 45 bpm were timed and reinforced using a metronome. The participants were instructed to copy the head movement to practice. During the illusory manipulation, the participants looked at their own video feed while listening to 45 bpm metronome beats in the background. They were instructed to make similar head movements in left-centre-right-centre directions in time with the metronome beats, as they did during the practice session. This was again performed for 90 seconds. The video feed of participants faces was viewed via the self-viewing feature of the video conferencing software Zoom 2. To create either a strong or weak enfacement illusion, the mirroring of the participant’s video feed was manipulated. In the strong (synchronous) enfacement illusion condition, the participants viewed a mirrored video feed in which the face on the screen moved in the same direction as they moved their head. In the weak (asynchronous) enfacement illusion condition the participants viewed an unmirrored video feed in which they saw their face on the screen move in the opposite direction to which they moved their head.
5. Measurement of subjective experience using enfacement illusion questionnaire – Following the implementation of enfacement illusion, participant’s susceptibility to the illusion was measured using a modified enfacement illusion questionnaire based on a previous enfacement illusion study 3. The questionnaire was a 12-item questionnaire that was completed by the participants immediately after each condition. Each questionnaire item was scored on a 7-point Likert scale, from strongly disagree to strongly agree. The subjective experience of face ownership was measure by items 1, 2, 3 and 4; face agency was measured by items 5, 7 and 8); the feeling of being younger by items 9 and 11; the feeling of being an adult by items 10 and 12; and there was control question (item 6); see table 2. The participants were also given an option to write about their subjective experience with a maximum limit of 100 words.
6. Autobiographical memory interview - A modified autobiographical memory interview (AMI) was conducted 4, 5. Soon after the participants completed the enfacement illusion questionnaire, they were once again shown their morphed or unmorphed face on the screen. While they continued to view their face, they were asked a set of questions about their autobiographical memories. The following instruction was given to the participants prior to each memory recollection 4 :
“I am going to ask you to tell me about an event from two time periods of your life. It can be either your childhood memory comprising of events that took place during your early childhood until the age of 11 years, or a recent memory which comprises of events that happened in previous one year. You will be given cue words ‘Home’ or ‘Holiday’ based on which you can choose any related event you wish. I will ask you to describe the events for a maximum of 5 minutes. I will then ask some questions about them to acquire further details, so be sure to only choose events that you feel comfortable discussing in detail. The event must be one you were personally involved in, and you must have a recollection of being personally involved. Do not pick events that you have heard about from others. They must be detailed events from a specific time and place. For example, simple recall of having played a basketball game would not be sufficient. However, an event involving a specific basketball game would be good. I want you to provide as much detail as you can about the event. Our interest is not in the specific characteristics of the event you choose, but rather the amount of detail you can remember about that event. So, I would again advise you to only pick events that you are comfortable sharing with me.”
For the AMI, there were 4 broad components – semantic recollection, free recall, recollection after general probing, and recollection after specific probing. The participants were asked to recollect either a childhood memory or a recent memory based on the cue words ‘home’ or ‘holiday’. There were three initial semantic (factual) questions asked. For example, if the memory to be recalled was for childhood holiday visit, the following was asked:
- Can you remember the name of a place you have visited When you were 11 years old or younger?
- Can you remember the year when the visit took place?
- Can you remember the full name or just the first name of a person who accompanied you, or that of a person you met during the visit?
After the semantic questions, the participants were asked to freely recall an episodic autobiographical memory of an event for 3-5 minutes. Following the free recollection, the participants were probed using general or non-specific questions such as ‘Is that everything you can recall about this incident?’. During this general probing phase, the interviewer also guided the participant to recollect just one event if they recollected multiple events during the free recollection. This was followed by specific probing phase during which the interviewer asked specific pre-structured questions to facilitate detailed recollection of event related episodic information related to time, place, episodic richness, sensory perceptions, and thoughts and emotions. Examples of specific probing:
- Can you further recall what happened during that [Main event]. For example, who else was there? What happened immediately before or afterwards? What was the weather like? How did you feel at the time? What was the behaviour or reactions of others around you?
- Can you recall in more detail the day, week or season, or the time of day at which the event (or parts of that event) occurred?
- Can you further recall the details of the location of the [Main event] for example, the town or city, street, building, room, or the outdoor location?
- Can you further recall the details of memory regarding what you heard, smelled, touched, tasted, saw, your body position (e.g.: were you sitting, lying down, standing etc,) and what was the duration of the [Main event]?
- Can you recall your emotional state and thoughts during the [Main event]?
7. Autobiographical memory interview scoring - Scoring of the AMI was split into 3 sections. The first section of scoring was for semantic questions. Successful recollection for each semantic question was awarded 1 point. If the participant recalled just the first name in the third semantic question, they were awarded 0.5 point for that question. Thus, each participant could score a maximum of 3 points for this section. The next section of scoring was for free recollection and recollection based on general probing. In the experiment by Levine, there was no significant difference in the scoring for free recollection and general probing therefore the memories recollected during these phases were scored together 4. Information was extracted based on the following six criteria – episodic richness, time, place, sensory perceptions, thoughts and emotions, and time integration. Episodic richness related to the qualitative estimate of reexperiencing. Time was scored based on the information recollected as it related to year, season, month, day of week and time of day. Place was scored for localisation of an event including, e.g., the city, street, building, room, and parts of room. Perceptual scores were based on auditory, olfactory, tactile, taste, visual details, body position and duration of the event. Scores for thoughts and emotions were given based on the emotional state and thoughts during the event. Time integration score was given for the event’s integration into a larger time scale as evidenced by inclusion of temporal contextual information or relation to other life periods. These six criteria were separately scored on a 4-point scale (0 – 3) as follows 4:
3 points – A rich, highly specific, evocative, and/or vivid description that appears to emerge from a feeling of reexperiencing.
2 points – A detailed description that falls short of 3 points in the degree of richness.
1 point – A description that is limited to general, nonspecific information but still episodic in nature.
0 points – No mention of information pertaining to the specified category, or a response that is based on semantic knowledge rather than episodic memory.
The third section for was for memory details recollected during the specific probing phase. Similar to the previously described section, this section was also scored on separate 4-point (0 – 3) scales for the 6 criteria – episodic richness, time, place, sensory perceptions, thoughts and emotions, and time integration.
8. Data analysis - Data collection from the four experiment conditions per participant resulted in a total of 200 memory interviews. Two researchers (co-authors), UG and PR separately scored the interviews for 10 randomly selected participants after blinding of the scores for different experimental conditions. Inter-rater reliability was assessed using separate interclass correlation coefficient (ICC) test for composite (average) scores of the 3 sections of AMI – i. Semantic scores; ii. Averaged scores for free recollection and general probing; and iii. Scores for specific probing. ICC estimates and their 95% confident intervals were calculated based on a mean-rating (k = 2), absolute-agreement, 2-way mixed-effects model. Once the acceptable ICC test results were achieved (values over 0.75) after randomly selecting data for 10 participants after each try, the interviews from the remaining 40 participants were scored by UG and PR (20 participants each) .
Data collection from the four experiment conditions per participant resulted in a total of 200 memory interviews. The statistical analysis for this experiment were performed using SPSS statistical package version 28.0.0.0 (SPSS Inc, Chicago, IL). Descriptive statistics were calculated and three MANOVAs in total were conducted to validate our illusory manipulation and Test our hypotheses.
The first MANOVA was conducted to examine the effect of group (child face filter, no face filter) and visuo-motor synchrony (synchronous, asynchronous) on enfacement scores - face ownership, face agency, child-like experience, adult-like experience, and control scores. The second MANOVA was performed to analyse the effect of group (child face filter, no face filter) and visuo-motor synchrony (synchronous, asynchronous) on childhood AM (100 memories) recollection scores – i. semantic, ii. free recollection and general probing, and iii. Specific probing based memory scores. The third MANOVA was performed to analyse the effect of group (child face filter, no face filter) and visuo-motor synchrony (synchronous, asynchronous) on recent AM recollection (100 memories) scores – i. semantic, ii. free recollection and general probing, and iii. Specific probing based memory scores. For multivariate statistical tests, normal and homogeneity of covariance were checked for any violation of assumptions using Shapiro-Wilk and Box’s M test. Multivariate analyses were additionally checked for multicollinearity. Univariate and multivariate Outliers were identified using box plots and by calculating the Mahalanobis distance and excluded from the analyses. An alpha level of 0.05 was implemented.
Correlations were ran to explore any trends between subjective enfacement experience (face ownership scores and face agency scores) and AM performance (composite semantic AM scores, composite episodic memory scores based on free recollection and general probing, and composite episodic memory scores based on specific probing). Correlations were ran separately for childhood AM and recent AM in both the groups. The normal distribution of variables and presence or absence of outliers were checked for any violations of correlation assumptions. Depending on the presence or absence of any violation of assumptions, parametric or non-parametric correlation tests were performed. An alpha level of 0.05 was implemented.