Title: The NIH Toolbox for Assessment of Neurological and Behavioral Function: The Cognition Team Experienc
1The NIH Toolbox for Assessment of Neurological
and Behavioral Function The Cognition Team
Experience
S. Weintraub, D. Tulksy, S. Dikmen, R. Heaton, R.
Havlik, N. Fox, D. Blitz, N. Chiaravalotti, P.
Zelazo, J. Machamer, N. Carlozzi, D. Mungas, C.
Nowinski, J. Slotkin, R. Gershon, B. Borosh
Center on Education Research and Outcomes,
Evanston Northwestern Healthcare, Evanston,
Illinois
- AIMS
- The NIH Toolbox, part of the NIH Neuroscience
Blueprint initiative, seeks to develop brief, yet
comprehensive, assessment tools measuring motor,
cognitive, sensory and emotional health and
function. Upon completion, the Toolbox will be
available for use in longitudinal epidemiologic
studies and prevention or intervention trials for
people ages 3-85. This is a preliminary report of
progress from the Cognition Team, including
identification of core constructs, selection of
candidate instruments and a plan for validation. - The mandate for all toolbox domain teams was to
- Cover 4-6 subdomains
- Take no more than 30 minutes to administer
- Cover the full range of normal function (not
disease states) - Cover ages 3-85 years
- Require minimal cost to the investigator
- Use state-of-the-art instruments and technology
- RESULTS
- 147 RESPONDENTS (63 OF INITIAL MAILINGS)
- 58 male
- Areas of expertise
- 57 adult only
- 18 pediatric only
- 25 adult and pediatric.
- Average years of research experience 25 (SD8)
- Average years of clinical experience 17
(SD13) - Experience conducting clinical research/trials
86 - Experience conducting longitudinal/epidemiologic
- research 87
- Results In-depth Phone Interviews
- 9 cognition experts (5 males, 4 females)
participated. - 6 reported expertise in neuropsychology.
Overlapping specialties cognitive neuroscience
(3), gerontology/geriatrics (2), psychometrics
(3), epidemiology (1), clinical trials (1),
neurology (2). - Average of 27.8 years (SD7.8) research and 24.2
( SD4.3) clinical experience, respectively. - Rankings of the sub domains previously identified
by the RFI and fine tuning of categories (I.e.,
breaking out working memory from executive
function and attention) were 1) Executive
Function, 2) Processing Speed, 3) Learning and
Memory, 4) Language, 5) Attention 6) Working
Memory, and 7) Visuospatial Functioning.
Additional domains mentioned as potentially
important included hot and cold cognition and
social cognition.
SAMPLE TASKS
Flanker Task Adaptation For Use With Children
And Adults
Flanker conditions can be used to measure
different aspects of attention and executive
function. The standard congruent condition
provides a measure of a simple continuous
performance type of attention, while the
standard incongruent condition minus the
congruent condition serves as a measure of
inhibition.
RESULTS INSTRUMENT SELECTION COMPUTER DRIVEN
CAT METHODOLOGY DIF ANALYSES
- METHODS
- Procedures
- Request For Information (RFI) And Expert
Interviews Online-surveys and subsequent
in-depth interviews of experts in cognition
and/or large cohort, pre-clinical and clinical
studies were obtained to identify criteria for
selection of Subdomains and measures. The
Scientific Advisory Board met to select final
criteria. - Instrument Selection Library search tabulation
of existing psychometric instruments and
measurement tools meeting with cognitive
consultants (July 27, 2007) to consider measures
and narrow options. - Instrument Development Subdomain teams disperse
to work on measures development team provides
technical support for computer administration
development plan. - Validation/Norming Plans Validation measures
discussed by the Cognition Team and in the July
27 meeting. Epi-Stats team designs sampling
plan computerized adaptive testing (CAT) and
Differential Item Functioning (DIF) Analyses.
Conceptual reasoning and set shifting are
measured with this task, designed for children,
but easily adaptable to adults.
- CONCLUSIONS
- A survey (Request for Information), in-depth
interviews, and a consensus meeting were used to
identify key constructs of Cognition and measures
for inclusion in Toolbox. Similar procedures
were used for the other three domains. Tests are
in development using computer technology.
Spanish version is planned. We anticipate the
final Toolbox to be available in 2011.
Ultimately, it can serve as a national resource
for the scientific community.
Acknowledgement Funded in whole or in part with
Federal funds from the National Institute on
Aging, National Institutes of Health, Contract
HHS-N-260-2006 00007-C. (PI Gershon) additional
contributors to this work include Jennifer Manly
PhD, Patricia Bauer PhD, Jennifer Richler PhD,
and Joanne Deocampo PhD.