MiroHealth
Follow the data
The brain is our most vital organ.
We help you treat it that way.
back to Our Science
PRELIMINARY TEST-RETEST
RELIABILITY AND LEARNING
EFFECTS
Miro’s preliminary reliability was investigated through
a test-retest reliability study that assessed
performance
in normal controls at three
time points over three months.
Demographics
DIAGNOSIS% F/MMEAN AGE (RANGE)TOTAL N=28
Normal controls83 / 1765.4 (49-89)21
High Functioning MCI70 / 3077.4 (52-95)7
(% F/M = percent female / percent male)
Participants and methods
Miro's test-retest reliability was evaluated
in 21 normal volunteers and 7 High
Functioning MCI participants
who were
assessed in their homes by a clinical
neuropsychologist on 3 occasions at
3-week intervals.
TIME POINT 1TIME POINT 2TIME POINT 3
50% of participantsMiro, ComparatorMiroMiro
50% of participantsComparator, MiroMiroMiro
Intra-class correlations were calculated
for the MCI Risk Score with data from 3
time points (T1, T2, T3) in order
to quantify
test-retest reliability, or the ability to
consistently identify performance levels
for specific
individuals. To assess learning
effects, the changes in scores across test
administrations for each subject
(trends)
were calculated. Permutation tests were
used to identify mean trends and their
significance.

The calculation of test-retest reliability
and learning effects included the
standardization of raw scores. Raw
scores
were standardized by the distribution of
scores in the normal subject population
upon initial Miro
assessment (T1). T1
observations for each score were
standardized with a mean value of 0 and
standard
deviation of 1. Subsequent
observations (T2, T3) were standardized
using the initial (T1) reference distribution.
Mean standardized scores in T2 or T3 are
measured in standard deviation units
relative to T1 scores. Slopes
have units of
standard deviations per assessment
(SD/A).
Results
The test-retest reliability intra-class
correlation coefficient (ICC) for the MCI
Risk Score was (0.79), with a 95%
confidence interval of (0.65, 0.89). This
shows the stability or reliability of
measurements of individuals’
functional
abilities.
Test-retest reliability

MODULEICCP-VAlUEOVERALL
MEAN SLOPE
TP 1 - TP 3
MEAN SLOPE
TP 1 - TP 2
MEAN SLOPE
TP 2 - TP 3
Digit Span Forward: longest span0.5490.004040.0400.081
Category Fluency: correct0.4610.0483-0.0548-0.09030.098
Verbal Learning: immediate recall0.6370.00504-0.0050.186-0.21
Digit Span Backward: longest span0.3860.001750.0880.455-0.28
Letter Fluency: correct0.4710.03960.08160.04780.252
Design Fluency: correct0.5530.0008370.1260.293-0.011
Trails B: time0.3950.002080.227-0.6640.726
Finger Tapping: number of taps 3rd trial0.5460.01350.240.0290.313
Coding: correct0.7253.39E-080.3220.3930.251
Discussion
Correlation between individuals’ scores
over time.
The intra-class correlation (ICC)
is the ratio of
inter-individual variance to
total variance of measurements across
time points. A narrow range of functional
ability as captured in this normal sample
set demonstrate a low inter-individual
variance relative to the
variance in a
mixed clinical population whose results
include scores outside the normal range.
Results show a
significant ICC, even
within this narrow range of normal
performance. This not only suggests
Miro’s ability to
consistently quantify
individual performance, but also Miro’s
ability to distinguish performance
signatures of
individual normal
participants.

Learning effects. Analysis of repeated
Miro assessments in a cognitively normal
population shows minimal
learning
effects or other trends over three
sequential test administrations. Across
Miro variables, the mean
subject
performance at time point 3 shifted less
than 0.09 standard deviations (SD) from
initial performance.
Follow the data
GET STARTED