The Guardian - 31.07.2019

Section:GDN 1N PaGe:3 Edition Date:190731 Edition:01 Zone: Sent at 30/7/2019 21:04 cYanmaGentaYellowbl

3

Wednesday 31 July 2019 The Guardian •

News

Ian Sample

Science editor

When Stephen Hawking wanted to

speak, he chose letters and words

from a synthesiser screen controlled

by twitches of a muscle in his cheek.

But the painstaking process the

cosmologist used might soon be bound

for the dustbin. With a radical new

approach, doctors have found a way

to extract a person’s speech directly

from their brain.

The breakthrough is the fi rst to

demonstrate how a person’s intention

to say specifi c words can be gleaned

from brain signals and turned into text

fast enough to keep pace with natural

conversation.

In its current form, the brain-

reading software works only for stock

sentences it has been trained on – but

scientists believe it is a stepping stone

towards a more powerful system that

can decode in real time the words a

person intends to say.

Doctors at the University of Califor-

nia in San Francisco (UCSF) took on

the challenge in the hope of creating a

product that allows paralysed people

to communicate more fl uidly than

using existing devices that pick up

eye movements and muscle twitches

to control a virtual keyboard.

“To date there is no speech pros-

thetic system that allows users to

have interactions on the rapid time-

scale of a human conversation,” said

Edward Chang, a neurosurgeon and

lead researcher on the study published

in the journal Nature.

The work, funded by Facebook,

was possible thanks to three epilepsy

patients who were about to have

neurosurgery for their condition.

Before their operations went ahead,

all three had a small patch of tiny

electrodes placed directly on the brain

for at least a week to map the origins

of their seizures.

During their stay in hospital, the

patients, all of whom could speak nor-

mally, agreed to take part in Chang’s

research. He used the electrodes

to record brain activity while each

patient was asked nine set questions

and asked to read a list of 24 potential

responses. With the recordings in

hand, Chang and his team built com-

puter models that learned to match

particular patterns of brain activity to

the questions the patients heard and

the answers they spoke. Once trained,

the software could identify almost

instantly, and from brain signals

alone, what question a patient heard

and what response they gave, with an

accuracy of 76% and 61% respectively.

“This is the fi rst time this approach

has been used to identify spoken

words and phrases,” said David Moses,

a researcher on the team.

“It’s important to keep in mind that

we achieved this using a very limited

vocabulary, but in future studies we

hope to increase the fl exibility as well

as the accuracy of what we can trans-

late.” Though rudimentary, the system

allowed patients to answer questions

about the music they liked; how well

they were feeling; whether their room

was too hot or cold, or too bright or

dark; and when they would like to be

checked on again.

Despite the breakthrough, there

are hurdles ahead. One challenge

is to improve the software so it can

translate brain signals into more varied

speech on the fl y. This will require

algorithms trained on a huge amount

of spoken language and corresponding

brain signal data, which may vary from

patient to patient.

Another goal is to read “imagined

speech” or sentences spoken in the

mind. At the moment, the system

detects brain signals sent to move the

lips, tongue, jaw and larynx – in other

words, the machinery of speech. But

for some patients these signals may

not suffi ce, and more sophisticated

ways of reading sentences in the brain

will be needed.

While the work is still in its infancy,

Winston Chiong , a neuroethicist at

UCSF who was not involved in the

latest study, said it was important to

debate the ethical issues such systems

might raise. For example, could a

“speech neuroprosthesis” unin-

tentionally reveal people’s private

thoughts?

Chang said decoding what someone

was trying to say was hard enough,

and that extracting their thoughts was

virtually impossible.

“I have no interest in developing

a technology to fi nd out what people

are thinking, even if it were possible,”

he said.

“But if someone wants to

communicate and can’t, I think we

have a responsibility as scientists

and clinicians to restore that most

fundamental human ability.”

Doctors hail breakthrough

in quest to read the mind

▼ The synthesiser used by Stephen

Hawking to communicate could be

replaced by a much faster system

PHOTOGRAPH: MURDO MACLEOD/THE GUARDIAN

New from John

Lewis: a panel

of burglars to

help keep

valuables safe

Rupert Neate

Forget hiding your valuables under

your pillow, in a safe or in your under-

wear drawer – those are among the

fi rst places that professional burglars

search for jewellery and cash.

A panel of former burglars, brought

together by the store chain John Lewis ,

advised holi daymakers that the safest

places for family heirlooms were cereal

boxes and among children’s toys.

One of the six former off enders said

cereal, pasta and rice packets made

the best hiding places because thieves

were “not going to go through all your

food packets”.

He said people often kept car keys

in the kitchen but they were usu-

ally easy to fi nd there. “Most people

have got a ‘bits and bobs’ cupboard in

their kitchen where they often keep

their keys,” he told John Lewis home

insurance.

Another former burglar advised

stashing valuables inside toy boxes

in children’s bedrooms, saying there

was an “unwritten rule” that thieves

did not enter youngsters’ rooms.

The panel, who were put in touch

with John Lewis via the St Giles Trust

charity, which helps former off end-

ers into legitimate work, said the rise

in online shopping ha d made it much

easier to identify unoccupied homes.

“ If you’re walking down the road

and see a parcel on a doorstep there

is a good chance someone is not in,”

one said. “I would suggest not ordering

parcels if you won’t be around when

they arrive .”

Leaving lights either on or off for

long periods was also a give away. The

best place to leave lights on when away

from home was the hallway, the panel

said. However, timer switches were

the best option for creating the illusion

that someone was at home.

Burglars can spend two months

watching a house before making

their move but as little as fi ve minutes

inside before fl eeing. Some of the for-

mer off enders said they took identity

documents, such as passports, driving

licences, credit cards and bank state-

ments, rather than TVs because of the

gains from selling on information.

Claire Nee, director of the Interna-

tional Centre for Research in Forensic

Psychology at Portsmouth University,

who interviewed hundreds of burglars

to analyse patterns in off ending, said:

“Identity documents are very valuable

at the minute due to identity theft for

fraud and people traffi cking.”

She advised being careful in con-

versation. “On the way to the airport,

talk loudly about your house sitter, not

about how you are looking forward to

your fortnight away.”

The Crime Survey for England and

Wales shows the number of domes-

tic burglaries fell by 2% to 669,

in 2018, compared with the previous

year.

How speech was decoded from the brain

Electrodes record brain

activity while participants

were asked questions and

read out answers

2

Electrodes placed on

participants' brains

1

Algorithm processes

brain activity for

questions and answers,

and learns to identify

what a participant hears

or says

3

Source: Nature

Frontal

cortex

Electrodes

Electrical pattern generated by

a listened-to question

Question

What musical instrument do you

like to listen to?

Correct answer

Electric guitar

Temporal

cortex

Accuracy rates for correctly

identifying the answer were 61%

РЕЛИЗ ПОДГОТОВИЛА ГРУППА "What's News" VK.COM/WSNWS

ЛИ

ЗП

ОДД

Дscale of a human conversation,” said scale of a human conversation,” said scale of a human conversation,” said scale of a human conversation,” said ГГГО

have interactions on the rapid time-

О

have interactions on the rapid time-

scale of a human conversation,” said scale of a human conversation,” said О

have interactions on the rapid time-have interactions on the rapid time-have interactions on the rapid time-have interactions on the rapid time-ТТО

В

thetic system that allows users to

В

thetic system that allows users to

have interactions on the rapid time-have interactions on the rapid time-В

thetic system that allows users to thetic system that allows users to ИИ

have interactions on the rapid time-

И

have interactions on the rapid time-

thetic system that allows users to thetic system that allows users to thetic system that allows users to thetic system that allows users to ЛАА

“To date there is no speech pros-“To date there is no speech pros-ГГ

thetic system that allows users to

Г

thetic system that allows users to

“To date there is no speech pros-“To date there is no speech pros-“To date there is no speech pros-“To date there is no speech pros-РУУП

to control a virtual keyboard.

П

to control a virtual keyboard.

“To date there is no speech pros-“To date there is no speech pros-П

to control a virtual keyboard.to control a virtual keyboard.to control a virtual keyboard.to control a virtual keyboard.ППАА

"What's "What's

scale of a human conversation,” said

"What's

scale of a human conversation,” said News"

thetic system that allows users to

News"

thetic system that allows users to

have interactions on the rapid time-

News"

have interactions on the rapid time-

scale of a human conversation,” said scale of a human conversation,” said News"

VK.COM/WSNWSVK.COM/WSNWS

have interactions on the rapid time-

VK.COM/WSNWS

have interactions on the rapid time-

scale of a human conversation,” said

VK.COM/WSNWS

scale of a human conversation,” said