Information Technology Reference
In-Depth Information
provision of real-time feedback remove the extra processing time required to operate
with the MAT (where stimuli are presented on the screen and the execution is
expected to happen on the dance mat). Alternatively, participants may have felt more
confident to step faster in space due to the absence of wearable sensors or a physical
apparatus.
Fig. 5.
Comparison between Average Reaction Times obtained with the Kinect and the MAT
per Participant per Panel
6
Conclusion and Future Work
This paper describes an evaluation where we assessed the ability of a Kinect-based
system to reliably perform the Choice Stepping Reaction Time (CSRT) task. This
study involves a technical assessment where the Kinect time-based measurements
were compared with a validated reaction time device that has proved to reliably
discriminate between fallers and non-fallers. The results showed high association
between the Reaction Times (RT) and Decision Times (DT) values obtained in both
systems. Movement Time (MT) however did not show agreement most likely for the
limitations in determining the initiation of the user's leg movement. Finally, RT
values were utilized to compute the CSRT test in order to assess the responsiveness of
the system. Statistical analysis showed that both systems correlate favorably,
suggesting that the Kinect system might be valid and reliable. More importantly, this
evaluation suggests that this Kinect-based system has the potential to be used in a
clinical setting as it can consistently measure both spatial and time based parameters
required for the fulfillment of a validated clinical test. Yet, the system's capability to
reliably discriminate between fallers and non-fallers still requires further evaluation.
The next stage of this study will involve the clinical validation of the step parameters,
followed by the conduction of controlled trials to determine the ability of the system
to discriminate between groups of fallers and non-fallers.
