Self-Referencing Bias in Psychiatric Inpatients

Overview

People tend to detect and recognise self-related information more quickly and efficiently than other kinds of information. For example, in a cocktail party, people are usually able to attend to just one conversation at a time. Messages from unattended conversations are rarely registered. However, most people would still hear their own name mentioned in unattended conversations. Research has shown that this self-referencing advantage manifests an individual's normal cognitive function and emotional wellbeing. It may be influenced by self worth and strength of self-esteem. Changes in self-related processing are hypothesised in various psychiatric conditions such as dissociative identity disorder and affective disorders, but the connection is poorly understood. Existing research mainly relies on self-report measures, which can be subjective and time consuming. This project will initiate a new approach that the investigators have developed to objectively measure self-related processing. The aim is to investigate how patients suffering from common psychiatric disorders respond to self-related information relative to age-matched control participants. The investigators also hope to establish whether the objective measurement of the self can form a new pathway to improve early diagnosis of mental health issues.

Objectives. The principal objective of this project is to find out whether common mental health problems result in altered self-referencing bias/advantage. The investigators will use a simple computer test to measure the strength of the self-referencing bias. The project will also explore whether common mental health problems could be partially indexed by deviations from the commonly observed self-referencing bias. If a deviation from the bias can form a reliable index, the investigators will also look at how the measure helps to harness other existing diagnostic methods for mental health issues. Scientific justification. People with various psychiatric disorders may have distorted and often negative views about themselves. For example, patients suffering from dissociative identity disorder are often unable to recognise their own identity. They may feel like a stranger to themselves. People diagnosed with depression or social anxiety may possess negative feelings about themselves. Even in a healthy sample, a negative mood or self-concept could be result in a reduced sensitivity to self-related information. However, little is known about the link between common mental health issues and self-concept. A recent study has provided the first empirical evidence about the link through a study of dissociative identity disorder. It was found that patients with this disorder are not more sensitive to the photo of self-face compared to a stranger's face. This deviates from the typical self-referencing advantage, because normal observers display greater sensitivity to self-related information such as self-face and self-name relative to information related to a stranger or friend. Affective disorders have also been linked to disturbances in self-definition. The study is interesting because it showed for the first time that a simple self-referential test could identify a deviation from the self-referencing bias in people with a psychiatric disorder, this could be used to index the dysfunction and recovery of this condition. It is an important finding in the field, because the link between the self-referencing bias and psychiatric disorders has been largely relying on more complicated interviewing techniques. However, as the authors of the study themselves acknowledge, a deviation from the common pattern of self-referencing bias may not be specific to dissociative identity disorder. Other psychiatric disorders may also result in an anomalous deviation from the typical self-referencing bias. The main objective of the present study is to identify the generality of the Lebois et al finding in other types of psychiatric disorders. Our hypothesis is that atypical self-concept is not specific to dissociative identity disorder, but underlying all common psychiatric disorders in general. Another important objective is to improve over their method, which has several crucial limitations that prevent an unequivocal interpretation. First, because their procedure presented a self-face or unfamiliar face for only 17 milliseconds, it demanded a very high level of concentration on the task to produce the expected self-referencing bias. However, it is difficult to rule out the possibility that due to ongoing mental-health issues, the participants in the patient group were less able to concentrate on the task. In fact, their data may support this interpretation as the patient group showed much less responses to images related to both self and others relative to the overall responses by the control group. It is therefore possible that their finding could be due to this extraneous factor rather than their patients being unable to discriminate a self-face from a stranger's face. Second, the self-reference bias in the control group could be explained by the familiarity with self-face relative to strangers' faces. However, the degree of the familiarity between faces of oneself and others may be different between the two groups. Unlike the control participants, patients with dissociative identity disorder might not be interested or equally interested in their own face hence they might look into mirrors or personal photos less frequently. If so, it may reduce the sensitivity of their detection of the self-face relative to other familiar faces, thus the difference between the two groups might be due to their different level of familiarity with self faces, rather than the lack of self-bias in the patient group. To rule out these alternative explanations, the proposed study will adopt a test first developed by the investigators. The test has been widely used over the past 10 years as an objective measure of self-bias and positive emotion. The test has gone beyond self-report measures typically used in psychiatry. The task completely eliminates the issues of differential familiarity. Instead of using familiar self-related information such as self-face and self-name, participants in this test are simply asked to learn some novel associations between some simple geometric shapes and labels representing different identities (Self, Friend, Stranger). For example, a participant could be told that a rectangle represents the participant, a circle represents the participant's friend, and a triangle represents a stranger. After the learning the associations, participants are shown either a correct shape-label paring (such as rectangle-participant and circle-friend) or incorrect pairing (such as circle-participant and triangle-friend). The participant's task is simply to make a simple 'yes' or 'no' response as quickly and as accurately as possible. Each pair is presented on a computer display until the participant makes a response. Thus the task does not require highly focused attention, which could potentially be an unreasonable demand for some patients. Also, unlike the test used in the aforementioned study that could be more difficult to implement because it requires taking a picture of the participant's face without their knowledge, our test may be a more feasible and reliable way of indexing the self-reference bias in patients with various psychiatric conditions. For normal participants tested in numerous studies, the typical finding of the matching test is a highly robust response time and accuracy advantage for judging the self-related pair relative to the others-related pairs. Self-referencing bias can be found in other tasks. For example, the investigators found previously that participants respond to self-face images more quickly than their friends' faces and the same effects are found in different cultures although the effects can be stronger in one culture than another. This measurement has been applied to a wide range of populations such as old adults and patients with brain lesions. Ageing, for example, can enhance response bias to friends, whereas visual extinction caused by parietal lobe lesion can leave the self-bias intact. The aim of this project is to apply this objective test for the first time to examine if the new measures can be used to track changes in self-representation in psychiatric disorders. This project comprises a novel collaboration between experimental psychology and psychiatry. The project will provide a proof-of-principle analysis of the clinical utility in various types of disorders. Our goal is to deepen understanding of the relationship between these disorders and abnormal pattern of self-concept. Hypothesis. The self-matching task in this project will provide the key answers to our research question. The main prediction of the study is that patients suffering from psychiatric disorders will show atypical self-referencing effect compared to results of the control group. This hypothesis is based on the theory that a coherent self-construal is the driving cause of the self-referencing effect. It requires a person to have a clear self-recognition and highly sensitive to self-referential information. In our self-matching task, this will be reflected by faster and more accurate responses to self-shape associations than other-shape associations. The effect in normal control is calculated by subtracting the mean of response time and response accuracy to self-referential shape from that to other-referential shape. The effect is usually about 5-10% based on the results in the past research. The investigators expect the effect to be lower than this in most of the patients. Method Participants. There will be two groups of participants, a patient group and a control group. Each group will have 60 participants. Patient group will consist of people admitted to an acute psychiatric ward. They will have a range of diagnoses comprising the spectrum of serious psychiatric disorders such as mood and anxiety disorders, psychotic disorders, and personality disorders. The control group will consist of age, gender, education matched participants who have no mental health issues. Materials and procedure. Participants will attend one test session. They will be tested individually and will be given two tasks. The first will require completing questionnaires of mood state. The second will be a computerised task on self-matching. Task 1: Questionnaires. To record the mood state, the investigators will use Positive and Negative Affect Schedule (PANAS-X) questionnaires (https://www.toolshero.com/psychology/personal-happiness/panas-scale/). The questionnaires contain two scales to measure a person's positive and negative emotions, respectively. The items on the questionnaires are 20 adjective words, such as 'Enthusiastic' (positive), 'Active' (positive), 'Guilty' (negative), and 'Upset' (negative). Participants are asked to indicate the extent to which each word describes how they have felt on the date of their participation based on a 5-point rating scale, where 1 = very slightly/not at all and 5 = extremely. This rating task takes less than five minutes to complete. The purpose of this task is to examine the effect of current mood on the result of the self-matching task. The investigators expect the low-mood to be associated with a lower self-referencing effect. This task will take about five minutes to complete. Task 2: Self matching. This is a computer-based task, which will be implemented after the questionnaires task. The purpose of this test is to measure self-bias. Participants will be taught to associate a geometric shape (such as a circle/square/triangle) with the self, and another shape with a close friend, and the third shape with a stranger. The assignment of the shape to a person will be counterbalanced across participants. These pairings will be shown on a computer display for participants to remember. After this initial learning stage, participants will be shown a series of a simultaneously presented pair of a shape and a label (self/friend/stranger). Half of these pairs are correctly matched, and the rest are mismatched. The order of matched or mismatched pairs will be random. The participants will be asked to judge on each trial whether the presented shape-label pair is a correct match. Participants will be asked to make a response as quickly and accurately as possible. A feedback message will be given for each response, which will inform the participant of whether the response was correct. Participants will be informed of their overall accuracy and response time at the end of each block. This is used to engage the participant for best task performance. The task will last about 10 minutes. Data analysis. The investigators will use (i) ANOVA to test the differences of self-related processing between the patient and control groups, and (ii) correlation analyses to assess the changes of self-representation in relation to the mood questionnaires and self-report measures.