Bedside EEG Test Aids Prognosis in Brain Injury Patients

Pauline Anderson

January 08, 2021

A simple, noninvasive electroencephalogram (EEG) may help detect residual cognition in unresponsive patients who have experienced a traumatic brain injury (TBI), results of a new study suggest.

The study showed that the use of a paradigm that measures the strength of responses to speech improved the accuracy of prognosis for these patients compared to prognoses made solely on the basis of standard clinical characteristics.

Dr Rodika Sokoliuk

"What we found is really compelling evidence" of the usefulness of the test, lead study author Rodika Sokoliuk, PhD, Center for Human Brain Health, School of Psychology, University of Birmingham, Birmingham, United Kingdom, told Medscape Medical News.

The passive measure of comprehension, which doesn't require any other response from the patient, can reduce uncertainty at a critical phase of decision making in the intensive care unit, said Sokoliuk.

The study was published online December 23 in Annals of Neurology.

Accurate, early prognostication is vital for efficient stratification of patients after a TBI, the authors write. This can often be achieved from patient behavior and CT at admission, but some patients continue to fail to obey commands after washout of sedation.

These patients pose a significant challenge for neurologic prognostication, they note. In these cases, clinicians and families must decide whether to "wait and see" or consider treatment withdrawal.

The authors note that a lack of command-following early in the postsedation period is associated with poor outcome, including vegetative state/unresponsive wakefulness syndrome (VS/UWS). This, they say, represents a "window of opportunity" for cessation of life-sustaining therapy at a time of considerable prognostic uncertainty.

Recent research shows that a significant proportion of unresponsive patients retain a level of cognition, and even consciousness, that isn't evident from their external behavior ― the so-called cognitive-motor dissociation.

The new study included 28 adult patients who had experienced a TBI and were admitted to the intensive care unit of the Queen Elizabeth Hospital, Birmingham, United Kingdom. The patients had a Glasgow Coma Scale motor score <6 (ie, they were incapable of obeying commands). They had been sedation free for 2 to 7 days.

For the paradigm, researchers constructed 288 English words using the male voice of the Apple synthesizer. The words required the same amount of time to be generated (320 ms) and were monosyllabic, so the rhythms of the sounds were the same.

The words were presented in a specific order: an adjective, then a noun, then a verb, then a noun. Two words ― for example, an adjective and noun ― "would build a meaningful phrase," and four words would build a sentence, said Sokoliuk.

The researchers built 72 of these four-word sentences. A trial comprised 12 of these sentences, resulting in a total of 864 four-word sentences.

Sokoliuk likened the paradigm to a rap song with a specific beat that is continually repeated. "Basically, we play 12 of these four-word sentences in a row, without any gaps," she said.

Each sentence was played to patients, in random order, a minimum of eight and a maximum of nine times per patient throughout the experiment. The patients' brain activity was recorded on EEG.

Sokoliuk noted that brain activity in healthy people only synchronizes with the rhythm of phrases and sentences when listeners consciously comprehend the speech. The researchers assessed the level of comprehension in the unresponsive patients by measuring the strength of this synchronicity or brain pattern.

After exclusions, 17 patients were available for outcome assessment 3 months post EEG, and 16 patients were available 6 months post EEG.

The analysis showed that outcome significantly correlated with the strength of patients' acute cortical tracking of phrases and sentences (r > .6; P < .007), quantified by intertrial phase coherence.

Linear regressions revealed that the strength of this comprehension response (beta = .603; P = .006) significantly improved the accuracy of prognoses relative to clinical characteristics alone, such as the Glasgow Coma Scale or CT grade.

Previous studies showed that if there is no understanding of the language used or if the subject is asleep, the brain doesn't have the "signature" of tracking phrases and sentences, so it doesn't have the synchronicity or the pattern of individuals with normal cognition, said Sokoliuk.

"You need a certain level of consciousness, and you need to understand the language, so your brain can actually track sentences or phrases," she said.

Sokoliuk explained that the paradigm shows that patients are understanding the sentences and are not just hearing them.

"It's not showing us that they only hear it, because there are no obvious gaps between the sentences; if there were gaps between sentences, it would probably only show that they hear it. It could be both, that they hear and understand it, but we wouldn't know."

A receiver operating characteristics analysis indicated 100% sensitivity and 80% specificity for a distinction between bad outcome (death, VS/UWS) and good outcome at 6 months.

"We could actually define a threshold of the tracking," said Sokoliuk. "Patients who had phrases and sentences tracking below this threshold had worse outcome than those whose tracking value was above this threshold."

The study illustrates that some posttraumatic patients who remain in an unresponsive state despite being sedation free may nevertheless comprehend speech.

The EEG paradigm approach, say the authors, may significantly reduce prognostic uncertainty in a critical phase of medical decision making.

It could also help clinicians make more appropriate decisions about whether or not to continue life-sustaining therapy and ensure more appropriate distribution of limited rehabilitation resources to patients most likely to benefit, the authors say.

Sokoliuk stressed that the paradigm could be used at the bedside soon after a brain injury. "The critical thing is, we can actually use it during the acute phase, which is very important for clinical decisions about life-sustaining methods, therapy, and long-term care."

The simple approach promises to be more accessible than, say, fMRI, said Sokoliuk. "Putting an unresponsive coma patient in a scanner is very difficult and also much more expensive," she said.

The next step, said Sokoliuk, is to repeat the study with a larger sample. "The number in the current study was quite small, and we can't say if the sensitivity of the paradigm is strong enough to use it as a standard prognostic tool," she said.

To use it in clinical setting, "we really have to have robust measures," she added.

She aims to conduct a collaborative study involving several institutions and more patients.

The research team plans to eventually build "an open-access toolbox" that would include the auditory streams to be played during EEG recordings and a program to analyze the data, said Sokoliuk.

"Then, in the end, you would get a threshold or a value of tracking for phrases and sentences, and this could then classify a patient to be in a good-outcome or in bad-outcome group," she said.

But she stressed this is a prognostic tool, not a diagnostic tool, and it should not be used in isolation. "It's important to know that no clinician should only use this paradigm to prognosticate a patient; our paradigm should be part of a bigger battery of tests," she said.

But it could go a long way toward helping families as well as physicians. "If they know that the patient would be better in 3 months' time, it's easier for them to decide what should come next," she said.

And it's heartening to know that when families talk to their unresponsive loved one, the patient understands them, she added.

Commenting on the study for Medscape Medical News, Christine Blume, PhD, Center for Chronobiology, University of Basel, Basel, Switzerland, whose research interests include cognitive processing of patients with disorders of consciousness, described it as "very elegant and appealing" and the paradigm it used as "really promising."

"However, we do of course not yet know about the prognostic value on a single-subject level, as the authors performed only group analyses," said Blume. "This will require more extensive and perhaps even multicenter studies."

It would also require developing a "solution" that "allows clinicians with limited time resources and perhaps lacking expert knowledge on the paradigm and the necessary analyses to apply the paradigm at bedside," said Blume.

She agreed that a passive paradigm that helps determine whether a patient consciously understands speech, without the need for further processing, "has the potential to really improve the diagnostic process and uncover covert consciousness."

One should bear in mind, though, that the paradigm "makes one essential assumption: that patients can understand speech," said Blume. "For example, an aphasic patient might not understand but still be conscious."

In this context, she added, "it's essential to note that while the presence of a response suggests consciousness, the absence of a response does not suggest the absence of consciousness."

Blume cautioned that the approach used in the study "is still at the stage of basic research." Although the paradigm is promising, "I do not think it is 'around the corner,' " she said.

The study was funded by the Medical Research Council. It was further supported by the National Institute for Health Research Surgical Reconstruction and Microbiology Research Center. Sokoliuk and Blume have disclosed no relevant financial relationships.

Ann Neurol. Published online December 23, 2020. Abstract

For more Medscape Neurology news, join us on Facebook and Twitter.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.