Computational protocol: Discourse intervention strategies in Alzheimer's disease:Eye-tracking and the effect of visual cues in conversation

Similar protocols

Protocol publication

[…] Participants. Participants comprised: (a) a control group of 10 older adults without dementia, who attended the Universidade Senior de Oeiras; and (b) five older adults with moderate-stage AD, who were outpatients of the Clínica São José located in Lisbon. The selection of participants with AD was performed by two neurologists of the Clínica São José (Lisbon, Portugal), who established the diagnosis of probable Alzheimer's dementia. (a) a control group of 10 older adults without dementia, who attended the Universidade Senior de Oeiras; and(b) five older adults with moderate-stage AD, who were outpatients of the Clínica São José located in Lisbon. The selection of participants with AD was performed by two neurologists of the Clínica São José (Lisbon, Portugal), who established the diagnosis of probable Alzheimer's dementia.The control group comprised predominantly women (70%); and, in the AD group, three out of five participants were women. The mean age of the control group was 78.31 years (6.65). In the AD group, the mean age was 80.92 (5.51) (U=89.00; p=0.35). In terms of educational level, there were no significant differences between groups (control: 6.12 (1.58); AD: 6.00 (1.66); U=108.00; p=0.88). The exclusion criteria of dementia for controls were: score ranging from 27 to 30 on the Mini-Mental State Examination (MMSE) (14; 15); autonomy in daily life; and absence of cognitive complaints, neurological, and psychiatric disorders based on data provided in an interview about health-related and sociodemographic data. The mean score of controls on the MMSE was 28.37 (1.02). The mean score of participants with AD on the MMSE differed significantly from the control sample (M=20.91; SD=4.25; p<0.05). Most of the participants with AD had moderate decline (MMSE=16). The study was approved by the Research Ethics Committee of Clínica São José in Portugal, where data collection took place. Participants included in the AD group and their relatives, as well as the control group, all signed a written consent form explaining the tasks planned for the study.Study design. This was a cross-sectional, quasi-experimental study, based on mixed analysis using comparison between groups, cases and controls, as well as comparison between conditions. Three different conditions were used: [A] conversations during which on-topic visual cues were displayed on a screen, [B] conversations during which off-topic visual cues were displayed on a screen, and [C] conversations during which a blank screen was displayed. All conditions were focused on the participants' personal reports about their youth years (between 20 and 30 years old). A family member of each participant previously selected two important events in the participant's life (one to be used in conditions A and B and the other for condition C). Family members were sent a confidential letter containing instructions on how to select pictures and create sentences to be used in the study. [A] conversations during which on-topic visual cues were displayed on a screen,[B] conversations during which off-topic visual cues were displayed on a screen, and[C] conversations during which a blank screen was displayed. All conditions were focused on the participants' personal reports about their youth years (between 20 and 30 years old). A family member of each participant previously selected two important events in the participant's life (one to be used in conditions A and B and the other for condition C). Family members were sent a confidential letter containing instructions on how to select pictures and create sentences to be used in the study.Youth years were established based on studies demonstrating special focus on autobiographical memory during this period, which is known as the memory bump for older adults. The pictures and sentences provided by family members were reviewed and selected to decide which event would be told during the experiment. The same event was used in conditions A and B to keep memory differences from interfering with the results of visual attention and discourse. Because of this, the events were used in a randomized order to prevent effects of practice from interfering with the results. Two researchers participated in the conversation experiment to ensure that the participants did not report the event twice to the same listener, which would increase context artificiality. A different important life event was used in condition C to avoid excessive effects of practice. With this in mind, each examiner's participation was balanced to avoid greater participation by any one of the examiners in a given condition, which could also interfere with the differences between the conditions. Discourse data of condition C were used to provide a means of comparison between the samples in a condition without visual cues.Pictures were chosen so that the image cues of conditions A and B were not significantly different in terms of those visual characteristics that tend to have an influence on the fixation time of pictures, such as number of people in the picture, size of the faces, presence of children and animals, and emotional arousal (pleasantness and facial expression arousal). Characteristics such as color, brightness, and size of the pictures were controlled using the computer program Photoshop by Microsoft, so that all characteristics were similar in conditions A and B. The sentences provided by family members were revised to be simple sentences containing approximately the same number of words and relevant information in order to be similar to the title of the event. The schedule of the experiment was set using the computer program E-prime.Procedures. A mobile, head-mounted eye tracker (SMI HED 50Hz incl. Polhemus head tracking) was used for data collection. Adjustments were made aimed at placing the camera and the lens in a position to capture appropriate images of the eye, including the analysis of corneal and pupillary reflexes on a computer used to record eye-tracking data. Describing the procedures prevented the AD participants from being afraid of wearing the reflective lens, which was positioned at a safe distance from the eye. Once an appropriate image was achieved, calibration was done.After calibration, the researcher pressed the key to start the presentation of the instructions on the screen, while the first examiner entered the room. The examiner sat facing the participant and explained the instructions that were being displayed on the screen:Next, you'll be asked to tell me about certain life events. Please tell the communicative partner sitting in front of you about the life event suggested. A picture and a sentence about the event might be displayed on the screen, or a picture and a sentence about a different event might be displayed, or the screen might be blank. Feel free to look at the screen for as long as you want while you are talking about the event. The next interviewer will not listen to your report. The second interviewer will substitute the first one and you will be asked to tell the same story again.Each participant was randomly exposed to the three conditions. In all the conditions, the picture and the sentence were displayed on the screen during the period of time each participant took to complete the story. Examiners were graduate students trained to limit their participation to certain speech acts during the conversation: (a) offering signs of attention, interest, and emotional reactions; (b) when participants interrupted their discourse, examiners provided verbal clues such as "What else"; and (c) when participants switched topic, the examiner provided a verbal clue to lead back to the topic. (a) offering signs of attention, interest, and emotional reactions;(b) when participants interrupted their discourse, examiners provided verbal clues such as "What else"; and(c) when participants switched topic, the examiner provided a verbal clue to lead back to the topic.Data analysis. Total fixation time, known as ‘dwell time' on regions of interest (picture, sentence, and interlocutor's face) was analyzed by checking possible differences between samples and between conditions.All discourse samples were transcribed verbatim and segmented into propositions. The coded variables were analyzed using the computer program CHAT of the CHILDS project. Discourse was divided into propositions and classified in terms of global coherence and informativeness according to a method adapted from Laine, Laakso, Vuorinen and Rinne. Twenty percent of the corpus was randomly selected and coded by two examiners, a speech therapist and a linguist blind to the conditions and to the group of participants. The Kappa test was used to evaluate the reliability of the discursive analysis. Global coherence showed 77% agreement; whereas informativeness had 81% agreement. According to Linell, Gustavsson and Juvonen, a 75% agreement is realistic for a conversation analysis, considering that interaction is often obscured by ambiguity. Carletta recommends that Kappa agreement should not be < 0.67 in order to allow reliable conclusions.Statistical analysis. Because of the objective characteristic of the eye-tracking data, small groups were compared using non parametric tests with the statistical package SPSS. The Mann-Whitney test was used to compare the groups, whereas the Wilcoxon test was applied to compare the visual tracking data between each of the conditions (within each sample). Regarding discursive data, each of the AD cases was compared with the control group to ensure a more accurate analysis. For this purpose, a special statistical program (SINGLIMS.EXE) was used for the comparison of cases with a small control group. According to this modified t-test, t values suggested significant differences between each AD case and the control group. The closer to zero the t values, the greater the likelihood of proving the null hypothesis that the case is part of the control group. […]

Pipeline specifications

Software tools E-Prime, SPSS
Applications Miscellaneous, Neuropsychology analysis
Organisms Homo sapiens