Electrophysiological markers of cognitive deficits in traumatic brain injury: a review.

Abstract

Event-related potentials (ERPs) and oscillatory activity from the human electroencephalogram (EEG) provides a rich source of data that helps elucidate specific processing impairments in TBI patients. This review will focus on some of the central and disabling cognitive deficits in TBI and how broadband ERP markers and the spectral content of the EEG can help explain abnormalities in brain function that impact upon processing speed, sustained attention, performance monitoring, inhibitory control and cognitive flexibility. Physiological signals also provide useful outcome markers in cognitive intervention studies in conjunction with behavioural endpoints. Potential rehabilitation approaches utilising electrophysiological markers of recovery are also discussed. Progress has been made in recent years in defining key pathophysiological mechanisms in the context of sensitive laboratory paradigms. However, aberrant physiological signals need to be understood more clearly in future studies in terms of the neuroanatomical impact of injury, particularly in relation to the most common type of damage in TBI, disrupting extended white matter fibres.

Full text

Page 1

??????????????????
Electrophysiological markers of cognitive deficits in Traumatic Brain Injury:
A review
Paul M. Dockree, Ian H. Robertson
PII:
DOI:
Reference:
S0167-8760(11)00005-5
doi: 10.1016/j.ijpsycho.2011.01.004
INTPSY 10250
To appear in:
International Journal of Psychophysiology
Received date:
Revised date:
Accepted date:
5 October 2010
22 December 2010
6 January 2011
Please cite this article as: Dockree, Paul M., Robertson, Ian H., Electrophysiological
markers of cognitive deficits in Traumatic Brain Injury: A review, International Journal
of Psychophysiology (2011), doi: 10.1016/j.ijpsycho.2011.01.004
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Page 2

ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Electrophysiological markers of cognitive deficits in
Traumatic Brain Injury: A review

Paul M. Dockree & Ian H. Robertson
School of Psychology and Trinity College Institute of Neuroscience,
Trinity College Dublin
Abstract
Event-related potentials (ERPs) and oscillatory activity from the human electroencephalogram
(EEG) provides a rich source of data that help elucidate specific processing impairments in TBI
patients. This review will focus on some of the central and disabling cognitive deficits in TBI
and how broadband ERPs markers and the spectral content of the EEG can help explain
abnormalities in brain function that impact upon processing speed, sustained attention,
performance monitoring, inhibitory control and cognitive flexibility. Physiological signals also
provide useful outcome markers in cognitive intervention studies in conjunction with
behavioural endpoints. Potential rehabilitation approaches utilising electrophysiological
markers of recovery are also discussed. Progress has been made in recent years in defining key
pathophysiological mechanisms in the context of sensitive laboratory paradigms. However,
aberrant physiological signals need to be understood more clearly in future studies in terms of
the neuroanatomical impact of injury, particularly in relation to the most common type of
damage in TBI, disrupting extended white matter fibres.

Introduction
Investigations using electrophysiological methods have developed our understanding of
damaged cognitive mechanisms in Traumatic Brain Injury (TBI). A wealth of analysis
techniques, sensitive to the timing of neural activity, has helped elucidate covert abnormalities
in brain function associated with overt behavioural deficits commonly seen after TBI. These
include behavioural changes in attention, memory, executive function as well as broader
personality and psychosocial changes. This review will explain how electrophysiological
markers can offer important insights into the nature of cognitive and behavioural impairments
seen at clinical assessment and throughout the course of rehabilitation.

Page 3

ACCEPTED MANUSCRIPT
This review is not an exhaustive account of all electrophysiological markers identified to date
ACCEPTED MANUSCRIPT
Neuroanatomical imaging techniques are evolving and researchers are gaining improved
anatomical resolution in the quantification of TBI-related tissue loss. However, in spite of this
progress, only generalised whole brain measures of volume loss consistently predict general
clinical outcome such as standard neuropsychological performance and information processing
speed. More specific brain-behaviour relationships have not been reliably identified perhaps, in
part, due to the distributed nature of the neuropathology (Levine, et al., 2006). The most
common type of damage in TBI is diffuse and occurs to the extended white matter (axonal)
fibres, particularly fronto-striatal, fronto-parietal and fronto-temporal pathways (Smith,
Meaney, & Shull, 2003). These lesions can result in a range of deficits, some highly specific such
as retrograde amnesia (Levine, et al., 1998) and others more general, such as those of the
attention-arousal system leading to impaired concentration and reduced alertness (Robertson,
Manly, Andrade, Baddeley, & Yiend, 1997). Recent evidence (Little, et al., 2010) points to some
specificity of damage at the loci of axonal fibres projecting to and from frontal cortices and
thalamic nuclei. Lesions to thalamo-cortical tracts account for considerable variance in
executive dysfunction in TBI. By contrast, damage to the cortical frontal lobe structures, cortico-
cortico tracts, or corpus callosum alone do not reliably account for these deficits. Although
findings such as these are encouraging in terms of understanding more central mechanisms that
underlie putative dysexecutive problems in TBI patients, the acquisition of physiological data
time-locked to specific behaviours offers a powerful approach that compliments anatomical
data. Event-related potentials (ERPs) and oscillatory activity from the human
electroencephalogram (EEG) provides a rich source of data that offers important insights into
cognitive processes in the human brain and provide a means of assessing neuropathology.
Specific brain systems can be dissociated by these electrical recordings and changes to the
component structure of ERP signals can lead to important inferences about the nature of specific
processing impairments.

but will focus of some of the core disabling cognitive deficits experienced by TBI patients and
how cognitive and electrophysiological recordings have elucidated intact and impaired
component processes during task performance. Moreover, the utility of electrophysiological
signals as an adjunct to cognitive intervention studies will also be discussed.
Processing Speed
One of the most generalised impairments following brain injury is reduced processing speed of
cognitive operations (Ferraro, 1996; Mathias, Beall, & Bigler, 2004; Mathias & Wheaton, 2007). It
has been demonstrated in a meta-analysis across 13 studies (Ferraro, 1996) that patients with
brain injury are 1.54 times slower than healthy controls on measures of simple and choice

Page 4

ACCEPTED MANUSCRIPT
effects of trauma. Specifically, the evidence suggests that qualitative disruption to early
ACCEPTED MANUSCRIPT
reaction time. Moreover, speed of information processing is also correlated with executive
functioning even when controlling for performance accuracy in TBI patients (Madigan, DeLuca,
Diamond, Tramontano, & Averill, 2000). Slowed processing speed on a range of reaction time
tests has been associated with the presence of diffuse axonal injury (DAI) following brain injury
and RT is disproportionately slower, in TBI patients than controls, especially on tasks requiring
the interhemispheric transfer of information where white matter integrity is essential
(Felmingham, Baguley, & Green, 2004).
From a clinical perspective, psychomotor function is commonly assessed using RT measures;
however, since a patient’s reaction time is the combined sum of both input (perceptual) and
output (motor execution) processes, it cannot be determined whether the basis of a processing
speed deficit is perceptual or psychomotor in origin or indeed both. Recording Event Related
Potentials (ERPs) that are both stimulus-locked and response-locked (Lew, Gray, & Poole, 2009)
circumvents this limitation and offers insight into the impaired neural processing underlying
slowed processing speed. Lew and colleagues analysed the RTs and ERPs of TBI patients and
controls that completed an oddball discrimination task. TBIs showed longer RTs to targets than
controls. Both a stimulus locked ERP component (P300) and a response locked ERP component
(Motor Potential, MP) were reduced in amplitude and delayed in latency in TBI patients
compared to controls suggesting that, in their TBI sample, both perceptual and psychomotor
processes were impaired.
The results of Lew et al. (2009) might imply that generalised slowing of both input and output
processes occurs in brain injury; however, under these circumstances it is difficult to discount
the possibility that the diffuse nature of TBI has disrupted the physical substrate of ERP
generation (i.e., reduced signal quality in general) rather than indexing the specific stages of
information processing speed deficits per se. Nevertheless, other research, utilising the precise
temporal resolution of ERPs, has revealed differential sensitivity of ERP components to the
perceptual discrimination processes has a downstream effect on the speed of processing by
delaying the transfer of information from stimulus processing to response selection. For
example, it has been reported (Reinvang, Nordby, & Nielsen, 2000) that early processing deficits
are apparent in discriminative stimulus processing and termination of irrelevant processing in
both active and passive oddball tasks (reflected by reduced N1 and P250 amplitude) in brain
injury patients compared to controls. Although, no latency differences were observed in these
early ERP waveforms, the later components in the ERP reflecting cognitive operations such as
stimulus-response mapping (the N2 and the P300) were significantly later in peak latency in TBI
patients compared to controls. This delayed neural processing may be a downstream result of
earlier perceptual deficits compromising discrimination of stimulus properties and rejection of

Page 5

ACCEPTED MANUSCRIPT
warning cues (increased P2, N2 and P3 latencies) as well as impaired deployment of attention to
the warning cue (reduced amplitude of P2) compared to controls. When the cue informed
ACCEPTED MANUSCRIPT
irrelevant information. Another study demonstrating qualitative disruption (Heinze, Munte,
Gobiet, Niemann, & Ruff, 1992) found that longer RTs and longer latency P3 responses were
apparent in patients relative to controls in a visual discrimination task. However,
disproportionate reductions in amplitudes were observed in intermediate ERP components (N1,
P2, and N2) compared to early (P1) and later (P300) waveforms implying differential sensitivity
of injury to discrete stages of processing. Specifically, the authors suggest that brain injury
impacts upon simple feature registration and early target discrimination as reflected by the
reduced intermediate components and this exerts downstream delays on neural processes
linked to stimulus-response mapping indexed by a prolonged P300 latencies in patients
compared to controls. However, since P300 latency mainly reflects stimulus-processing time,
in contrast to response-processing, its sensitivity as index of any response-time delays in TBI
patients is best seen in fast conditions, with its sensitivity decreasing when response times get
longer.
A more direct way to assess processing speed deficits in TBI patients that manifest themselves
between stimulus decoding and response preparation has come from studies investigating the
Contingent Negative Variation (CNV) waveform. The CNV is elicited in tasks that present a
warning stimulus (e.g. a tone or visual cue) which forewarns the presentation of an upcoming
target that requires a response (e.g. withhold/respond). The morphology of the CNV is first
observed after the warning stimulus in which a series of positive and negative polarity shifts
over fronto-central scalp are evoked followed by in a pronounced negative signal described as
the early CNV. This is followed by a sustained negativity that persists until the presentation of
the target – the late CNV. Consistently, TBI patients show longer RTs to the target stimulus
than controls (Campbell, Suffield, & Deacon, 1990; Cremona-Meteyard & Geffen, 1994;
Segalowitz, Dywan, & Unsal, 1997; Segalowitz, Unsal, & Dywan, 1992). In one of the latter
studies (Cremona-Meteyard & Geffen, 1994) participants showed delayed processing of
participants to withhold to the subsequent target, patients showed persisting activation of the
late CNV suggesting perseverative response activation. In another study examining CNV
abnormalities in TBI patients (Rugg, et al., 1989) the early CNV following the cue did not
differentiate go and no-go trials suggesting a deficit in orientation to the cue information and,
again, no attenuation of the late CNV when a no-go stimulus was upcoming. The evidence
suggests that the longer response times in TBI patients may occur because of impaired cue
processing that, in turn, disrupts communication with responses-end systems. Furthermore,
healthy controls show an association between greater amplitude of the late CNV and shortened
RTs suggesting the late CNV is a good index of the efficiency of which the response system is
activated in the healthy brain. By contrast, TBI patients, in the abovementioned studies