Objective Assessment of Behavioral Associations of Patients With Dementia

1. Objective(s): The key goal of this project is to develop methods for objectively
monitoring and, ultimately, predicting outbursts of aberrant behaviors commonly seen
among patients with a variety of neuropsychiatric conditions: dementia, PTSD, TBI, and
schizophrenia. This initial study will be on patients with dementia. The analytic focus
is on the measurement of electrodermal activity (EDA), a widely-studied physiological
recording that indicates sympathetic tone; the central hypothesis in this approach is
that sympathetic tone will increase prior to an outburst and, therefore, may allow
therapeutic interventions to prevent it. Alterations of sympathetic activity have been
implicated in the literature for each of these four conditions. Here, EDA recording
holds the advantages of requiring neither effort nor communication on the part of the
subject, and of being a noninvasive procedure. These factors make EDA potentially useful
for the clinical assessment of arousal among patients with dementia, which will be the
first of the specific groups studied under this protocol.

2. Research Design: The general design of all of these studies is the monitoring of EDA in
subjects with a small device that can be worn on the wrist, ankle or arm; patterns of
spontaneous EDA will then be assessed for features that may predict the occurrence of an
outburst. In the initial study of dementia patients, the investigators will first survey
EDA responses in commonly used paradigms (e.g., visual stimulus, EDA response).
Currently, little is known about EDA in this population, and the investigators redress
this by (1) establishing behavioral and physiological baselines for each subject in the
study, (2) determining how these two classes of dependent measures cohere, and (3)
evaluating whether patterns of EDA differ as a function of aberrant behaviors that are
often seen in dementia: wandering, combativeness, and "sundowning" (agitation that
typically develops late in the day). These behaviors are often grouped in the literature
under the term "disorders of arousal", and are clinically well known to present as signs
of disorientation and in some cases, distress. (4) Moreover, quantitative assessment of
EDA change conceivably may reveal patterns that may signal the onset of problem
behaviors.

Each subject will serve as his/her own control, providing within-subject measures of EDA
signal over time and its correlation with aberrant behaviors. It is worth noting that
the key EDA measures are not subject to bias based on experimenter ratings. Further, the
aberrant behaviors will also be rated over time, using quantitative rating scales for
each of them. In a second class of analyses, the investigators will perform group
contrasts of subjects with and without these aberrant behaviors, as operationally
defined by rating scales, to determine if patients with these problem behaviors are
distinctive as a group with respect to EDA patterns. All statistical analysis will be
performed on anonymized data, blind to group membership.

3. Methodology: EDA will be measured by standard techniques in all subjects: electrodes may
be placed on the hands, feet, ankles, wrists, or arm according to preference and
configured to measure conductivity or potential. In one portion of the study, a wearable
device to measure EDA will be used to determine whether EDA predicts the occurrence of
an outburst. In the study with dementia patients, assessments will be performed at an
inpatient dementia care facility at the Bedford VAMC (GRECC). Over a two-year period,
the investigators anticipate enrolling 60 subjects. Recruitment criteria include
toleration of the EDA recording device worn on the wrist (similar to a watch) or the
ankle. Skin potential, skin conductance, and membrane electrical power will be monitored
and data recorded electronically. Data will be de-identified by assigning the serial
subject number to each patient; a table linking the number to the subject's identity
will be kept on a secure VA server, allowing re-identification after data analysis.

4. Findings: If successful, EDA may predict an outburst of aberrant behavior with enough
lead time to permit a preventive intervention.

5. Clinical Relationships: The investigators hypothesize that: (a) stable physiologic
baselines can be obtained for EDA measures, (b) EDA signals are altered as a function of
aberrant behaviors ("disorders of arousal") on a within-subject basis, and (c) group
differences will emerge between samples of patients with and without these aberrant
behaviors.

6. Impact/Significance: If the EDA were borne out as a sensitive measure of disorders of
arousal, this measure may find application in the assessment and treatment of dementia.

Inclusion Criteria:

- any patient admitted to dementia unit at Bedford VAMC

Exclusion Criteria:

- inability to cooperate with electrodermal measurements