Technology in Motion (TIM) - Project Holocue
Towards on-demand and assist-as-needed patient-tailored cues to alleviate and prevent freezing of gait

Freezing of gait (FOG) is one of the most disabling motor symptoms of
Parkinson*s disease (PD) and a major cause of falls with a debilitating impact
on quality of life. There is a general consensus among clinicians that cues can
be an effective therapeutic to help alleviate FOG episodes once they have
occurred. Whereas recent studies suggest that 3D cues may be more effective
than 2D cues, patients* responses to specific cueing modalities vary strongly,
calling for an individually tailored approach for presenting the right type(s)
of cues. With Holocue, we therefore aim to alleviate FOG by presenting
patient-tailored cues on demand, such as horizontal bars to step over. We have
built the Holocue application for Microsoft Hololens, an untethered
non-occluding mixed-reality headset with a holographic display unit through
which 3D holograms can be displayed in one*s own environment. By focusing on
alleviating FOG, Holocue holds the potential to improve treatment beyond
current standards of care. It may be particularly helpful for PD patients with
FOG in the dopaminergic *ON state*, for whom currently no further
evidence-based medication is available to alleviate these disabling motor
symptoms.

The primary objective of this study is to explore the potential efficacy of the
Holocue application for alleviating FOG episodes through on-demand cue
activation. In order to achieve this, walking data obtained from the Hololens
will first be validated against walking data obtained from a reference motion
capture system in healthy control subjects (Hololens validation study).
Subsequently, the usability and acceptability of the Holocue application will
be examined in PD patients with FOG in the dopaminergic *ON state* in a
laboratory setting (Holocue efficacy studies 1 & 2) and in the patient's home
environment (Holocue efficacy study 2). Using the data of Holocue efficacy
study 2 we will develop a generic classification algorithm to predict FOG from
Hololens data (movement and/or environmental data) for assist-as-needed cues at
the right time and location, which will be validated in an independent group of
PD patients (i.e., technical validation study for accurate and timely FOG
prediction). In addition, the generic classification algorithm will be
patient-tailored to determine if this improves the prediction of FOG.

Observational, cross-sectional study in which walking data is collected in
healthy controls and PD patients.

Validation study (healthy controls; 1 session; movement laboratory): Overground
walking data of the Hololens will be systematically collected and validated
against the Interactive Walkway, a reference motion capture system. Subjects
will walk over the 8 m-walkway at different imposed step lengths, walking
speeds and cadences by means of presented visual context or auditory cues.
Total measuring time (including instructions and practice) will be
approximately 60 min.

Efficacy study 1 (PD patients, movement laboratory, 1 session): The Interactive
Walkway will be used to elicit FOG by means of suddenly projecting visual
obstacles, stop-and-go cues and narrow turning indicators on the 8 m-walkway,
all with and without cognitive dual tasking to vary the likelihood of FOG
occurrences. All tasks will be performed while wearing the Hololens (collecting
movement data of normal walking and during pre-FOG and FOG episodes), with and
without the on-demand Holocue functionality to establish its potential efficacy
in alleviating FOG episodes once they have occurred. FOG will also be assessed
in real-world situations in and around the lab that are known to elicit
freezing (e.g., turning, walking through narrow passages). Total measuring time
(including instructions and practice) is approximately 90 min.

Holocue Efficacy study 2 (PD patients, movement laboratory/home, 3 sessions):
Session 1 is a home-based session supervised by a research assistant. Patients
will walk in their own home environment, indoors and outdoors, and are
encouraged to visit *freeze prone* locations. Patients will walk these routes
with and without wearing the HoloLens (without Holocue functionality). Session
2 is a laboratory session, during which patients will be accustomed to the
Holocue application and trained in selecting and activating cues, and during
which the potential efficacy for alleviating FOG will be explored in the
laboratory setting (three conditions: Holocue, HoloLens without Holocue,
Control without HoloLens. NB: improvements implemented based on Holocue
efficacy study 1). Session 3 is again a home-based session. Patients will walk
similar routes in and around their house as in Session 1, supervised by the
research assistant, while wearing the HoloLens with and without the Holocue
on-demand application. During all sessions, PD patients will wear two
shoe-mounted inertial measurement units.

Technical validation study (PD patients, home, 2 sessions): This study includes
the external validation of the generic classification algorithm developed using
the data of Holocue efficacy study 2 to detected FOG episodes (i.e., timing and
accuracy). In addition, the generic classification algorithm will be
patient-tailored to determine if this improves the prediction of FOG. We will
recruit 12 new patients suffering from FOG using the same criteria of efficacy
studies 1 and 2 to obtain truly independent test data for external validation
of the generic classification algorithm. This study will be conducted in
participants* free-living environments, consisting of two non-standardized
supervised walking sessions in and around the participant*s home as part of
their daily routine (e.g., doing groceries, visiting friends, leisure walking).
Participants will be encouraged to visit freeze-prone locations while wearing
the Hololens to collect environmental and movement data. The research assistant
will always walk close to the patient for safety reasons. Per 150 minutes
visit, we aim for 20-40 minutes of walking. During all sessions, PD patients
will wear two shoe-mounted and two trunk-mounted IMUs in order to be able to
cross-validate the classification algorithms.

All trials of Holocue efficacy studies 1 and 2 and the technical validation
study will be filmed and will be annotated by two independent reviewers (e.g.
FOG episode, walking, voluntary stop) to determine 1) the efficacy of the
Holocue application and 2) if FOG episodes can be validly detected and
predicted using movement data and/or environment data of the Hololens (for
future development of preventive cueing, i.e., to present cues before a FOG
episode occurs).

The risks of the assessment are minimal. The markerless tracking devices of the
Interactive Walkway comprise sensors consisting of an infrared laser
transmitter and an infrared camera, which allows for unobtrusive assessment of
walking (i.e., no direct contact with the subjects, no potentially harmful
radiation, and no cumbersome marker placement). The Hololens, which will be
used for presentation of virtual cues to alleviate FOG episodes, allows for
unlimited vision of the actual environment. Both systems (i.e., the markerless
tracking devices of the Interactive Walkway and the Hololens) are commercially
available and have been successfully used in various areas of application and
target groups.

Actual measurements will take no longer than 60-150 minutes per session (1 to 3
sessions per participant), including 30 minutes (in total) for clinical testing
in PD patients. Walking will be done at a comfortable walking speed. If
necessary, PD patients are allowed to use their walking aid. The experimenter
can walk along with the patient for assurance or to provide support. Additional
rest periods will be offered if necessary. For PD patients, if possible the
measurement in the *Technology in Motion* lab will be combined with a regular
visit to the outpatient clinic, such that no extra travelling is required.
Travel costs to the hospital will be compensated on the basis of public
transport (2nd class) or travelled distance by car (¤ 0,19 per km) plus
compensation for parking costs. Participants in the Holocue efficacy study 2
and the technical validation study will also receive a VVV voucher of ¤20 and
¤40, respectively.

In summary, the risks of this study are minimal and the burden on the
participants is low. On the other hand, the benefits of this study are expected
to be great. Holocue addresses one of the most disabling motor symptoms of PD,
namely that of FOG, which prevalence increases from about 10% in Hoehn and Yahr
stage 1 to more than 90% in the advanced stage 4. By focusing on alleviating
and ultimately preventing FOG, Holocue holds the potential to improve treatment
beyond current standards of care. It can be particularly helpful for PD
patients with FOG in the dopaminergic *ON state* (i.e., often present in
atypical parkinsonian disorders and in the late stages of PD), for whom
currently no further evidence-based medication is available to alleviate these
disabling motor symptoms.