Cau­tious re­sponse to US re­search into pre­dic­tion of risk

The Irish Times - Tuesday - Health - - Health Suicide - SEÁN DUKE

Amer­i­can re­search into the pre­dic­tion of the risk of sui­cide in at-risk in­di­vid­u­als has pro­voked a cau­tious re­sponse from Ir­ish ex­perts.

The study, pub­lished in Na­ture Hu­man Be­hav­iour, sug­gested cal­cu­lat­ing the risk of sui­cide could be sub­stan­tially im­proved through brain scan­ning tech­niques such as func­tional mag­netic res­o­nance imag­ing (fMRI).

It also sug­gested that ma­chines with ar­ti­fi­cial in­tel­li­gence could pick up neu­ral signs of sui­cide risk through brain-scan anal­y­sis.

“There is op­ti­mism in our field that this ap­proach can be use­ful in the fu­ture, but we’re not there yet,” said Prof Gary Dono­hoe, School of Psy­chol­ogy, NUI Gal­way, who is also part of the team at the Neu­roimag­ing and Cog­ni­tive Ge­nomics Cen­tre at NUIG work­ing on an in­ter­na­tional re­search project to bet­ter pre­dict the course of diseases such as schizophre­nia.

“That will only hap­pen when these kinds of an­a­lyt­i­cal ap­proaches can pre­dict not just what di­ag­nos­tic group you are cur­rently in, but what group you are likely to be­long to in the fu­ture.”

The re­searchers, led by Marcel Just, pro­fes­sor of psy­chol­ogy at Carnegie Mel­lon Univer­sity in Pitts­burgh, mea­sured the re­sponse in peo­ple’s brains to words with neg­a­tive con­no­ta­tions, such as death, as well as words such as good, which are con­sid­ered pos­i­tive.

The emo­tional re­ac­tion to such words was al­tered in the 17 peo­ple stud­ied who had ideas of sui­cide, com­pared to 17 healthy in­di­vid­u­als. “For ex­am­ple, death evokes more sad­ness and shame,” said Prof Just.

“Our tech­nique has the ca­pa­bil­ity of not only de­tect­ing that cer­tain thoughts are al­tered in peo­ple with sui­ci­dal ideation, but it also pro­vides in­for­ma­tion about the na­ture of the al­ter­ation. This in­for­ma­tion pro­vides a tar­get for treat­ment, which could aim to re­duce or elim­i­nate the al­ter­ation.”

Prof Just said his team used a ma­chine learn­ing al­go­rithm (soft­ware de­signed to learn and get clev­erer from ex­pe­ri­ence) to iden­tify peo­ple with sui­ci­dal ideas based on their fMRI scans.

The pre­dic­tion of sui­cide risk is no­to­ri­ously dif­fi­cult. There would be great in­ter­est in any new tech­nique that could re­li­ably help clin­i­cians to im­prove risk assess­ment in ad­vance of dis­charg­ing a pa­tient, for ex­am­ple.

“This study does not add to the ac­cu­racy of risk assess­ment,” Brendan Kelly, pro­fes­sor of psy­chi­a­try at Trin­ity Col­lege Dublin, cau­tioned, “although it might point to a use­ful av­enue”.

The assess­ment of the risk of sui­cide over the course of a life­time varies be­tween 0.5 per cent and 1.2 per cent, said Prof Kelly.

For a per­son in Ire­land to­day who lives to be 90, the sta­tis­ti­cal risk of sui­cide is 0.93 per cent, he said. That fig­ure rises sig­nif­i­cantly to 10-15 per cent for in­di­vid­u­als who have ma­jor de­pres­sion.

Prof Kelly sounded a fur­ther note of cau­tion, say­ing the num­ber of peo­ple that took part in the Carnegie Mel­lon study was very small.

The only way to defini­tively show that us­ing fMRI, or other brain scan­ning tech­niques to pre­dict sui­cide risk worked, would be to scan peo­ple’s brains, make a risk pre­dic­tion, and fol­low them up to see whether they did, in fact, die by sui­cide, said Prof Kelly.

Newspapers in English

Newspapers from Ireland

© PressReader. All rights reserved.