Highlights from AAIC 2023
Donanemab trial results
The headline from this year’s AAIC in Amsterdam is undoubtedly the announcement of the results of TRAILBLAZER-ALZ-2, a phase 3 trial of Eli Lilly’s new drug Donanemab with 1,800 participants. Donanemab is an antibody-based drug that works by breaking down already-formed amyloid plaques that build up in Alzheimer’s disease (AD), rather than interfering with their production.
The results presented from this trial are promising as they showed significant slowing in cognitive and functional decline associated with AD compared to the placebo group, as measured on two scales – the iADRS and the CDR. Although clearing plaques alone may not be sufficient to prevent decline, it was encouraging to see that the difference of trajectories between placebo and treatment arms continued even after withdrawing treatment, with the suggestion that short interventions may be of value.
Increasing interest in automated and remote cognitive assessments, including voice
With my cognitive hat on, I notice increasing interest in use of cognitive assessments that can be delivered automatically and remotely (we noted widespread work evaluating language in picture description tasks!), and a strong focus on repeatable tasks that may be sensitive to prodromal MCI and AD. My WinterLight colleague Melisa Gumus noted considerable interest at her poster in the use of language assessments to explore different neuropsychiatric symptoms (such as anxiety) in AD and other dementias. I was enthused to see posters continuing to explore the mnemonic properties of word stimuli in health and AD-type pathologies, with exciting implications for task sensitivity to disease.
Adapting Paired Associates Learning (PAL) for mobile
I presented a set of posters on our work adapting PAL for a smartphone format. These show that there is strong comparability between the classic, web-based version of PAL and our new smartphone edition, and explored their use interchangeably, meaning that researchers have good reason to consider smartphone PAL in remote screening use cases, and as a key component of remote testing batteries.
- Download: Comparing the Classic Paired Associates Learning Task with a new Smartphone-Based Version of the Task
- Download: Exploring the Behaviour, Reliability, and Usability of a new Smartphone Version of Paired Associates Learning
Developing a high-frequency testing version of the Verbal Paired Associates (VPA) task in a Learning Over Repeated Exposures (LORE) paradigm
The second pair of posters we presented covers our preliminary testing of VPA-LORE and the very encouraging early results that have led us to the wider-range testing we are now undertaking.
The first poster demonstrates that the VPA task and parallel sets of word-pair-lists we have developed for it are well suited to use in the low-burden, high-frequency LORE task. This is exciting to us, as LORE has already been shown to be sensitive to prodromal Tau pathology and Amyloid load .
In this task requiring only two minutes’ interaction with our NeuroVocalix system each day for a five-day ‘burst’, users hear the same eight word-pairs each day and try to remember them. Over five days their recall of the set of word-pairs gradually improves, giving us a measure of their rate of learning. The poster shows how our VPA LORE task captures three important features of memory function:
- immediate recall on the first day
- learning retention during an overnight delay on day two
- learning over repeated exposures during days two to five.
In our preliminary testing, the learning shown by a sample of healthy older adults covers a good range of memory ability, a crucial feature for successful high-frequency longitudinal testing. This work paved the way for the next poster that showed a remarkable consistency between two five-day bursts of learning separated by 9 months in our age 55+ participants. This consistency is another critical feature for repeated testing and long-term tracking of learning ability.
- Download: Feasibility of Remote Automated Verbal Paired Associate (VPA) Learning Over Repeated Exposure (LORE)
- Download: Reliability of VPA Learning Over Repeat Exposures Across an Extended Interval Using Automated Remote Method
AAIC this year has once again provided me with fantastic opportunities for information exchange with leading scientists in both academia and industry. As always, it is tremendously valuable to us at Cambridge Cognition to meet in person with our existing and future friends, clients, and collaborators, and we look forward very much to the next instalment.
Access all our AAIC 2023 content here