About



Contact Information

[back to top]
Chloé Pou-Prom, MSc Candidate
Principal Investigator
Bahen Centre for Information Technology
40 St George St, Toronto, Room BA4202
Toronto, Ontario M5S 2E4
514-710-9994
chloe@cs.toronto.edu
Maria Yancheva, MSc
Co-Investigator
D. L. Pratt Bldg
6 King's College Road, Room PT276 D3
Toronto, Ontario M5S 3G8
416-830-2434
yancheva@cs.toronto.edu
Frank Rudzicz, PhD
Faculty Supervisor
Toronto Rehabilitation Institute - UHN
550 University Avenue, Room 12-175
Toronto, Ontario M5G 2A2
416-597-3422 x7971
frank@cs.toronto.edu

Overview of study

[back to top]

The Computational Linguistics group at the Department of Computer Science in the University of Toronto is conducting a study to analyze the differences in longitudinal progression of linguistic markers between healthy older adults and ones with dementia. We are interested in assessing the validity and usability of an automated web-based and phone-based system (Talk2Me) for linguistic data acquisition, as means for longitudinal monitoring of changes in language ability.

Purpose – Why are we doing this study?

[back to top]

The number of people affected by dementia grows every year, and research into the pathogenesis and progression of neurodegenerative disorders becomes ever more important. There is evidence of pathophysiological changes, including biomarker changes and subtle cognitive decline, occurring decades before clinical diagnosis. Cognitive abilities which may be impacted include memory, response time, visuo-spatial orientation, and language production and comprehension.

We are interested in investigating how features of language, such as syntactic complexity, lexical richness, semantic content, and acoustics change over time, in populations of older adults who are (a) healthy, (b) asymptomatic but later develop dementia, and (c) diagnosed with dementia. We would like to determine whether automatic analysis of language measures is as, or more, effective at predicting pathology than in-person tests of cognitive ability performed in a clinical setting.

Among those already diagnosed with dementia of the Alzheimer's type, we are interested in finding out whether linguistic decline occurs in different areas among different subgroups. For instance, are some individuals with clinical dementia more likely to experience lexical changes (e.g., vocabulary impoverishment), while others to experience loss of syntactic complexity? Such symptomatic differences could signify a different path of physiological progression of the disease.

To conduct these analyses, we need a large corpus of linguistic data from people in various age groups, both asymptomatic (with no existing diagnosis) and symptomatic (with clinical diagnosis of dementia). Since subtle changes in language are salient markers of cognitive decline, we need multiple data samples for each individual, collected at periodic intervals (e.g., weekly, monthly or annually). The collected dataset will be a large step in the direction of preclinical assessment of cognitive decline, which can eventually lead to early diagnosis and treatment of neurodegeneration.

Procedures – What will happen during the study?

[back to top]

Step 1: Informed consent

If you choose to participate, you will begin by creating a user account and logging into this website. Upon first logon, you will be shown a consent form, and a link to this help page. Your informed consent, or that of an alternate decision maker who can provide consent on your behalf, is required in order to participate; you can provide your consent by selecting the appropriate option at the bottom of the consent form. The contact information of the research investigators will also be provided; you can request additional information about the study and ask any questions prior to providing consent, as well as at any time after that.

Step 2: Demographic survey

Upon consenting to participate, you will be asked to complete a short demographic survey. The survey is mandatory for participation, since any linguistic analyses of the data need to control for variables such as age, level of education, pre-existing diagnosis of dementia, language proficiency, etc. All data collected here will only be analyzed and reported anonymously (i.e., your username and (optional) email address will not be disclosed, while demographic data will be reported in aggregate only).

Step 3: Language data acquisition

Sessions will either be completed over the website or over the phone.

You can complete as many sessions as you want – the more, the better. You can opt in for scheduled reminder emails which will prompt you to complete a new session on a weekly or monthly basis. Your email will remain fully confidential: it will not be used for any other purpose, it will never be provided to third parties, and it will never be publicized.

No training is needed to use the system, but if you experience any difficulties or have any questions, you can contact the research investigators at any time. You can also choose to withdraw from the study at any time, by selecting the option from your dashboard, or by contacting the researchers. If you choose to withdraw, all data collected from you will be deleted immediately and irreversibly, and you will receive confirmation.

Step 4: Usability survey (optional)

If you have completed at least one full session on the system, you will be given the option to complete a usability survey. The survey will help us determine if the website is easy to use, and if there are any issues such as unclear instructions, difficulty locating information on the site, or legibility issues. If you ever felt frustrated using the system, this is a great way to send us your feedback – we will take it into account and try to improve the interface.

Compensation and acknowledgement

If you have completed at least one full session on the system, you will receive a personalized Certificate of Participation (the name you provide for the certificate will not be stored anywhere; your data remains fully confidential, and all analyses and publications – anonymous).

Follow-up updates

If you are interested in the research conducted with the data collected during this study, you can optionally opt in to receive updates of publications via email. Updates will not be sent more frequently than once monthly. Your email will remain fully confidential: it will not be used for any other purpose, it will never be provided to third parties, and it will never be publicized. You can opt out of these updates at any time, by selecting the "unsubscribe" option in the emails you receive, or selecting the appropriate option from the dashboard.

Risks – Are there any bad things about the study?

[back to top]

There are no known research risks associated with the study. If you choose to participate, one session may take a total of 15 minutes at a time. You will only need to use standard hardware peripherals (a keyboard, a mouse, a microphone, and a telephone).

Participants with probable and/or diagnosed dementia may be exposed to psychological and emotional risks, such as feelings of confusion or anxiety as a result of participation. If you feel confused, frustrated, anxious, or stressed during any part of the study, you can stop at any time by exiting the active session, or by completely withdrawing from the study. If you withdraw, all of the collected data will be deleted immediately and irreversibly.

Benefits – Are there any good things about the study?

[back to top]

A recent recommendation of the National Institute on Aging and Alzheimer's Association emphasizes the importance of the early identification of preclinical asymptomatic stages of dementia. Early detection of pathophysiological processes leading to AD can make disease therapy more effective, leading to reduction or elimination of symptom progression.

The dataset collected in this study will be a large step in the direction of preclinical assessment of cognitive decline, as it will enable research in the patterns of longitudinal progression of linguistic changes in large groups of symptomatic and asymptomatic adults. Analyses of differences in linguistic feature progression have the potential to form the basis for early detection of neurodegeneration.

Confidentiality – Who will know what I did in the study?

[back to top]

Data stored in the database will be identified with a unique randomly-generated but non-identifying numerical code (ID). General demographic data (e.g., age, gender, level of education), as well as linguistic data collected during the study, will be associated with each ID. Optionally collected data, such as email address, will be kept completely confidential and never reported.

All data obtained in this study will remain as private as possible. Your username and optionally provided email address will remain fully confidential, and will not be used when reporting results. Analyses based on the collected data will report linguistic and demographic data only in aggregate. Linguistic and demographic data will not be shared with anyone, unless you explicitly consent to making it publically available for other researchers. It should be noted that linguistic data, such as audio recordings collected during the study, may be used to discover one’s identity.

The data transfer from your computer to the secure web server takes place over a secure communication protocol. The server is located on university premises, behind locked doors. Only the three investigators listed on this page (Dr. Frank Rudzicz, Chloé Pou-Prom, and Maria Yancheva) have access to the data. The data is stored in a secure database with encryption. If and only if you provide explicit consent for doing so, we may share the data with other researchers in the future; your username and email address will never be shared with anyone.

Subject's Rights – What are the decisions that I can make about this study?

[back to top]

You can ask any questions at any time and they will be answered by the investigators, whose contact information is listed on this page.

Participation in this study is entirely voluntary, and it is your choice whether to participate or not. If you agree to participate, you have the right to withdraw at any time, without giving a reason. If you choose to withdraw, all of your data recorded as part of the study will be deleted immediately and irreversibly. There will be no consequences.

You can optionally choose to:

You can opt out of any of the above selections at any time. You can opt in and out any number of times, with no consequences. Your email will remain fully confidential: it will not be used for any other purpose, it will never be provided to third parties, and it will never be publicized.

If you require any additional information, please contact the research investigators.