Testing for dementia among the elderly could one day be as simple as talking into a smartphone thanks to speech-analysing technology being developed by engineers at UNSW Sydney.
About the Tool
An app that uses machine learning technology will reportedly look at paralinguistic features of a person’s speech as well as test memory recall.
Paralinguistic features include prosody, pitch, volume and intonation.
The Engineering Researcher, who is leading the development of new algorithms that will be deployed on a population-wide scale using a smartphone app, is from the University’s School of Electrical Engineering and Telecommunications.
Dr Beena Ahmed explained that the tool will essentially replace current subjective, time-consuming procedures that have limited diagnostic accuracy.
Current Screening Methods
- Current screening for dementia in older adults involves structured interviews and tests assessing their ability to perform various mental activities.
- Two commonly used assessments are the Mini-Mental State Exam (MMSE) and the Mini-Cog test.
- In the MMSE, a clinician asks a patient a series of questions designed to test a range of everyday mental skills. In the Mini-Cog, meanwhile, a person is asked to complete two tasks.
- The first is to memorise three common objects and be able to recall them a few minutes later.
- The second is to draw the face of a clock showing all 12 numbers in the right places showing a time specified by the examiner.
- The researcher explained that accurately scoring assessments like these is subjective, depending on the expertise of the clinician, test conditions, patient condition and other factors.
- In addition, access to specialist services is challenging and often inequitable, especially in rural and remote areas.
- Speech, on the other hand, is easy to collect, even remotely over the phone, is cost-effective and non-invasive.
- Since the app will perform the speech-based assessment computationally, it is less susceptible to subjective biases.
How it Works
To use the app, a participant listens to a list of 15 words and is then asked to repeat back as many as they can remember.
This is done three times using the same 15 words. Then after a gap of 20 minutes, the person is asked to recall as many of the words as they can.
Apart from the accurate recall of the words, the app will be listening for tell-tale signs of dementia.
These include frequent pauses in searching for a word, repeated or restarted phrases, repeated or extended syllables, frequent fillers such as “um”, repaired utterances, mispronunciations, word substitutions as well as certain effects in the speaker’s melody, intonation and rhythm.
Initial studies by the team have shown that it is possible to discriminate between participants at high risk and low risk of dementia as determined by clinicians with an accuracy of 94.7% when trained with paralinguistic features only.
Meanwhile, there is an accuracy of 97.2% when trained with paralinguistic and episodic memory features using audio recordings of participants completing an episodic memory test.
Benefits of the App
The main goal of the screening tool is to help identify individuals at risk of cognitive decline so they can be provided with treatment to delay the onset of dementia.
Practitioners could use the results to direct those at high likelihood of cognitive impairment to primary care for further assessment and care as well as link users to validated primary and secondary prevention tools.
The results could also be used to contribute to accurate longitudinal cognitive data while also reducing primary care workload by alleviating the necessity of in-person cognitive screening.
Other advantages that the app could bring include a reduction of referral time to specialist clinics (where accessible) as well as supporting large scale cognitive trials at low cost.
The development of the app is in its early stages and could be available for use in five to 10 years.
