Neurophysiological
assessment/Pathology
characterization
The MORE analysis methodology facilitates a view of
functional brain activity providing detail to understand system-level
compensation and recovery. It provides a full statistical distribution of
brain function describing the characteristics of either repeated
experimental trials or those of multiple participants in a study. Since
MORE analysis provides a description of the activities of multiple areas of
the brain, it facilitates representation of these activities as a
multi-dimensional statistical distribution of relationships among active brain
areas. Hence, the interactions of all identified parts of the brain that
are involved for a given human behaviour are represented. Through this
representation, the brain function of individual persons can be compared to
others in the group. Hence, one can see how brain area activation and
brain area coordination in individual persons compare to others in the
group. This type of comparison can be used to reveal not only possible
brain dysfunction but also how other areas of the brain (that are atypically
used in the healthy population) are being used to compensate for injury.
This multi-dimensional perspective, where each participant for each trial
in a single participant study is a data point, allows for the application
of automated clustering algorithms to identify groupings within the
distribution of data. Hence, classification of pathologies and
identification of stub-types of these pathologies is supported.
The development of computer games for diagnostics
and rehabilitation
The MORE analysis methodology is suitable be used to
design a class of video games known as ‘serious games’. Serious games are
those that are not designed strictly for entertainment uses but have been
optimized for learning, training, or rehabilitation purposes. The algorithms
and methodology can be used to better understand the video game experience
by showing how aspects of the video game interact with the human brain.
Information acquired in this way can help designers make better video games
and understand how people solve a variety of challenging problems.
Evaluation of drug treatments and therapies
Because MORE analysis provides a sensitive measure of
the activities of specific areas of the brain and how these brain areas
interact with each other, subtle changes in brain function can be visible.
The MORE analysis algorithm is more sensitive than other EEG methods
because of how it rejects noise in the EEG and ‘zooms in’ on the activities
of specific areas of the brain. This might make MORE analysis effective
for detecting small changes in brain function related to drug treatments
and therefore make it suitable to facilitate the evaluation of new
pharmaceuticals and how they affect various aspects of human brain
function. For example, MORE analysis might be effective to (a) maximize
efficacy of a drug treatment, (b) identify possible negative behavioural
and cognitive effects of a drug before the effects are strongly evident in
behaviour, (c) provide an objective means to measure cognitive changes
related to drug treatment, (d) identify new treatment applications for new
drugs.
The methodology can potentially be used to maximize the
efficacy of drug treatment because sub-types of the broader classification
of a pathology, say-- Parkinson’s disease, can be identified based on
subtle differences in the brain function of the Parkinson’s disease
population and how they respond to low doses of a drug. Multiple benefits
are achieved when a sub-type of a population is identified that most
benefit from a particular drug treatment. First, the overall measure of
efficacy that is used by regulatory agencies to determine the usefulness of
a drug improves because best-matched persons would be receiving the drug
during clinical trials. Second, necessary variations in drug formulation
and dosage to maximize patient benefits and minimize side-effects might be
identified. This is achievable because small changes in brain function
might be identified that a patient receiving the drug may not notice or be
able to articulate. Small changes in brain function should be visible for
small dosages and thus negative behavioural side-effects can be mitigated
before they become a problem when prescribed large dosages are
administered. Third, identification of sub-type might help identify why
particular drugs are not effective in some cases. Fourth, and perhaps the
most interesting use of the MORE analysis methodology is how it can be used
to identify new markets for existing drugs. This application is possible
because MORE analysis can be used to inform drug researchers about what
areas of the brain are affected by the drug. Hence, an objective empirical
measure can be obtained that describes the function of brain areas
important in various behavioural and cognitive domains (fine limb movement,
task switching, anxiety).
Web marketing (Continue Where Eye-Tracking Leaves Off)
MORE analysis and the MOST-EEG algorithm can also be used to help in optimizing web site and media content. Many of us have witnessed the recent popularity of using eye-tracking to understand how people interact with web sites. Eye-tracking research has provided us objective information to guide page lay-out. However, eye-tracking does not tell us the whole story. While it tells us where a user's eye is pointing, it does not tell us where in their visual field their attention is directed. It does not tell us how, or if, their brain is processing the content provided. The MOST-EEG technology may be useful in providing this function.
|
Relation of new technology to current
methods
Researchers often use
scalp-acquired EEG methods to obtain non-invasive measures of brain
activity to study the relationship between brain function and human
behaviour. The EEG is a very weak representation of electrical activity
occurring in the brain. Because it is so weak and is essentially a mixture
of all of the activities of all active brain areas with various sources of
interference, it is difficult to discern meaningful information from the
EEG. In turn, this makes it is difficult to use the EEG to examine complex
brain function.
The standard approach to EEG
analysis requires that data be obtained from many people in an experiment
and used to create a picture of average EEG activity. This means that
comparisons can be made only between groups of people (e.g., children
versus adults) or between two experimental conditions (e.g., healthy versus
brain-injured) instead of between individual people. This also means that
standard methods strive to create simple pictures of the EEG (by averaging)
and not pictures of brain activity. Another weakness of this approach is that
it ignores differences between individuals who are grouped together,
assuming everyone in a group is basically the same. Analogously, it is
like drawing an average face -- the detail of individual differences is
lost. The result is that standard analysis cannot be used to tell if an
individual has unusual brain activity. Often it is the unusual activities
that are interesting and can add to our knowledge and understanding.
Standard EEG methods also
require that the brain activity behind the EEG be as simple as possible in
order to be able to make sense of the EEG data. Imaging having to analyze
a recording of 100 simultaneous conversations to try and pick out what one
specific person was saying. It would be easier to discern what this person
was saying if there was only one conversation. Similarly, researchers try
to limit the number of brain areas that are active while recording EEG data
by simplifying the behaviour during EEG recording. Researchers generally
only examine EEG gathered while study participants make very
straightforward yes/no decisions about sensory stimuli in order to activate
as few brain areas as possible. Unfortunately this limits the complexity
of the cognition that can be studied and the amount we can learn about
brain-behaviour relations.
The primary objective of my
research was to create an automated algorithm and general analysis
methodology for mining information from the EEG. The analysis methodology
compares the results of the algorithm’s EEG analysis from several people
and experimental conditions to determine which brain areas are commonly
active in most people, and which brain areas might not be functioning
properly in particular individuals.
One of the big problems in
the field of EEG analysis is known as the ‘inverse problem’. This problem
states that unless you know the number of active brain areas that
contribute to a given pattern of EEG, there are an infinite number of
possible locations of active brain areas that could have produced it.
Simply, unless you know the number of contributors, you cannot know where
they are. The major advance of my research was to develop an algorithm for
identifying the number of unique active brain areas that contribute to the
EEG and an algorithm to isolate their waveform and topography (spatial
distribution on the scalp) before finding their locations in the head.
Once success of the analysis
algorithm had been demonstrated, the feasibility of
using it to analyse EEG collected during a cognitively complex task
was investigated. EEG data collected from a small number (12) of study
participants was examined. These data were collected as participants navigated in two different virtual
environments (levels in the UnReal video gaming environment), each requiring a
different kind of cognition.
Even with this small number of participants, using a
relatively inexpensive low-resolution EEG apparatus, I was able to report
several interesting results. First, I was able to identify a large number
of brain areas involved in spatial navigation. These included areas
responsible for seeing the environment, areas responsible for recognizing
objects, and areas responsible for understanding the spatial relations
between them. Second, I was able to tell which areas of the brain had
coordinative activities, possibility communicating with each other.
Finally, I found indications that the analysis can distinguish individuals
who different from the others, suggesting that the procedure might be
useful in medical diagnostic situations.
|