Life Integrated Technology 3: Speech Restoration

Need of the Technology

Paralysed from head to toe, the patient, his mind intact, is imprisoned inside his own body,
unable to speak or move. One can’t imagine the means of communication in such a critical
condition. But in case of this patient in the past years, his left eye was the only source of
communication. New technologies based on brain­machine interfaces could have helped to
escape the imprisonment— and not just through flights of imagination.
The question that comes to the mind is how is it possible?

Introduction

During the past decade, researchers have developed an electrode array that feeds a digital
camera’s output into the visual cortex of a blind person’s brain. Arguably the most ambitious of
today’s brain­machine devices is one that aims specifically to help locked­in patients
(​characterized by full consciousness with near­complete paralysis​)​ by converting their inner
thoughts to real­time synthesized speech. But the effort to develop this electronic “speech
prosthesis” shows just how difficult it can be to meld mind with metal.

For people who have lost the ability to talk, a new "phonetic speech engine" from Illinois­based
Ambient Corporation provides an audible voice. Developed in conjunction with Texas
Instruments, the device uses electrodes to detect neuronal signals travelling from the brain to the
vocal cords. Patients imagine slowly sounding out words; then the quarter­size device (located in
a neck brace) wirelessly transmits those impulses to a computer or a cell phone, which produces
speech.

Fig 1: Speech Restoration Device

Implementation

The first step in the development process was verification and fine­tuning of the raw signals, the
next step was gathering and processing of data. In this process, various sounds were played and
the patient was asked to try to repeat them aloud in his mind while neural recordings were taken
via the electrode. The next hurdle was the interpretation of the noisy firings of neurons as
intelligible speech.
Then the pattern of neuron firings has to be translated into appropriate sounds which require
some very sophisticated signals processing techniques. Later it was noticed that the job is not to
decode the neuron firings perfectly, but to do it in a way that allows the patient to “tune” the
prosthesis quickly as it plays back the sounds which the patient is trying to vocalize. After that,
the researchers extend the range of sounds to other vowels and consonants.

Fig 2: Phonetic Alphabet

Still a long ahead

The process implemented using the algorithms based on “formant frequencies” to make up the
spoken sounds leads to the development of other prosthetic systems, for a wider range of
paralyzed patients. Later it was suggested to multiple cone electrodes in different locations in the
motor cortex, for example, rather than using a large array. Other groups have favoured electrode
arrays to gather signals from a greater number of neurons, but most electrode arrays have proved
to be relatively short­lived.

Fig 3: Schematic Representation of Speech Restorer

Conclusion

In the near term, the goal is just to get the speech prosthesis technology into as many locked­in
patients as possible. There are no hard data on the prevalence of the condition, but there is a
belief that there are probably thousands of such people are kept at nursing homes and are cared
by their spouses or parents.

Thus this study points out the evolution from the initial devices to newly emerging ones which
are now available such that the patient is now able to select from a variety of different devices
each with its own advantages and disadvantages which however would add to the
armamentarium in obtaining fluent speech from the impaired patient.

Aakash Chhatlani