Iris Murdoch is a famous and highly regarded author, who not only made a valuable contribution to the world of literature, but also (inadvertently) to dementia research.
Like any gifted writer, Murdoch drew upon an impressive use of syntax and vocabulary range in her work, but later in her career – when her language ability should have been peaking – her linguistic faculties began to fade.
Professor Peter Garrard – a consultant neurologist at St George’s University Hospital, London – hypothesised that the decline in her use of novel words may have been a “pre-symptomatic marker” of dementia – a condition she later went on to be diagnosed with.
“I noticed that Murdoch’s vocabulary and syntax was increasingly simplistic in some of her later work – quite unlike her earlier novels,” Prof. Garrard told Informa ahead of the National Dementia Conference 2020.
“We have known for quite some time that Alzheimer’s affects a person’s use of language, but the question was whether these language difficulties may emerge far sooner than the onset of the condition – and therefore be a pre-symptomatic marker of the pathology.
“It was quite plausible that the early presence of amyloid – which we know to be a precursor of dementia – could cause language problems in pre-dementia.”
Indeed, Garrard’s research later confirmed a correlation between a decline in language faculties like syntax and vocabulary with a later diagnosis of dementia. This was supported by a number of follow up studies.
The research used a complex algorithm to identify mean frequency counts (per million) of novel and general usage words and showed that Murdoch was introducing far fewer novel words in her later novels than in her earlier work.
“What we saw in Murdoch’s final novel was that her vocabulary was less varied, with the mean frequency count of novel words she used markedly lower than in her first book. This was a hint that she was using much simpler syntactic structures,” said Garrard.
“This allowed us to conclude that people living with Alzheimer’s tend to lose words from their vocabulary that are low in frequency in general usage.”
The finding was significant in that it meant researchers could identify dementia, long before symptoms presented – a point in which the patient may be less receptive to treatment.
“By the time dementia symptoms, like memory loss and confusion, start to present it’s often too late,” said Garrard.
“These findings will allow us to intervene much earlier and significantly increase the person’s chances of offsetting the condition.”
The language test would also prove to be far more cost-effective than other diagnostic methods, such as brain imaging for the presence of amyloid – a precursor to tau (the neuro-proton linked to cognitive destruction in dementia).
“The diagnosis could be made cheaply at huge scale and at a time of disease progression when a patient is more likely to respond to treatment,” said Garrard.
A reliable method
The findings have subsequently been demonstrated in more complex models of language structure – and in shorter samples of speech than were analysed in Murdoch’s books.
The hypothesis also holds in general members of the public, not just those with an aptitude for writing.
“It has moved from a headline grabbing one-off study to the brink of becoming a useable and practical technology for clinical use,” said Garrard.
“We’re excited about the possibility of it making it’s way into mainstream practice.”
Professor Peter Garrard is among a stellar line-up of international and domestic speakers to present at the National Dementia Conference on 13-14 May 2020. Hear more from him on the groundbreaking research and its implications for mainstream practice.
Learn more and register.