Developing and validating a finger tapping test for Parkinson’s disease

‘The Neurology group conducts many studies in patients with PD,’ says Dr Robert-Jan Doll, Associate Director and head of the BEA group. ‘So they were seeking an objective measurement of motor symptoms that would be straightforward to apply and could be frequently repeated.’ Such a measure would provide a more versatile alternative to the Movement Disorder Society’s Unified Parkinson’s Disease Rating Scale Part III (MDS-UPDRS III), which requires systematic evaluation by a clinician. The team decided to develop a finger tapping task, in which the patient is asked to press two buttons alternately. Parameters relating to the speed, accuracy and rhythm of taps can yield insight into the severity of motor symptoms.

The first prototype used physical buttons. However, this approach did not produce sufficiently consistent results. ‘Pressing the buttons takes some force, which comes at the expense of speed and accuracy,’ says Jeremy Cicilia, Clinical Research Engineer in the BEA group. The team therefore developed an alternative method using a touchscreen. ‘An additional advantage of touchscreens is that you can collect more forms of data — for example, the number of taps that miss the target,’ says Cicilia.

[text continues below the photo]

Initial testing in healthy volunteers

Soma Makai-Bölöni, Clinical Scientist in the BEA group, was responsible for the first exploratory study using touchscreens. ‘Various finger tapping tasks can be found in the literature, but they are often only described in general terms,’ says Makai-Bölöni. He conducted a study with 14 healthy volunteers to map out the different possibilities and determine which to take forward in patient studies.

Makai-Bölöni: ‘We looked at alternate finger tapping (AFT), in which the patient taps with their index and middle finger, and alternate side tapping (AST), where the index finger is used to tap targets which are further apart. We tested both options with and without visual cueing.’ In terms of endpoints, the initial study revealed that the total number of taps and spatial error had the highest repeatability and sensitivity. Meanwhile, a simpler setup without visual cueing is preferred, to maximise the ability of the test to detect medication effects.

As part of this initial phase, it was also necessary to describe the technicalities of the setup in detail. Abdes Maatoug, Validation Engineer in the BEA group: ‘To use a task in clinical drug development, it has to be very precisely defined in terms of standard operating procedures and technical validation procedures. We work through each procedure to verify functionality before a method or tool is clinically validated.’

Placebo-controlled trial in PD patients

The next step in the development was a study in 20 PD patients. Eva Thijssen, Experienced Clinical Scientist: ‘To ascertain whether finger tapping is suitable for measuring drug effects, we did a placebo-controlled crossover study with the most commonly used Parkinson’s medication: a combination of levodopa and carbidopa.’ Patients participating in this study were given placebo or levodopa/carbidopa after a one-night withdrawal from their own PD medication.

AFT and AST were then applied without visual cues. A thumb-index finger tapping task was used as an additional source of data, with a goniometer to measure index finger flexion. An MDS-UPDRS III score was recorded pre-dose and up to three-and-a-half hours post-dose. Thijssen: ‘Based on the data, we can conclude that the AST and thumb-index finger tapping tasks provide short, rater-independent measurements that are sensitive to dopaminergic drug effects, with an effect size similar to the MDS-UPDRS III.’

Composite biomarker

The data from this study were then further analysed by Ahnjili Zhuparris, Data Scientist in the BEA group. Zhuparris: ‘The end goal was to have an objective and repeatable measurement of PD motor symptoms which is sensitive to drug effects. To this end, we applied machine learning to develop a composite finger tapping biomarker.’ A composite biomarker would reflect co-occurring manifestations of motor function impairment in PD patients, and could subsequently be used to monitor or predict differential longitudinal treatment effects. Zhuparris’ work demonstrates that the AST composite biomarker may detect treatment effects earlier and more accurately than the MDS-UPDRS III. ‘This provides evidence in favour of adopting the AST as a composite biomarker for monitoring anti-Parkinsonian drug effects in clinical trials,’ says Zhuparris.