Conceived and designed the experiments: PA JC. Performed the experiments: PA JC. Analyzed the data: PA JC. Contributed reagents/materials/analysis tools: JC. Wrote the paper: PA JC.
The authors have declared that no competing interests exist.
Freezing of gait (FOG) in Parkinson's disease (PD) rises in prevalence when the effect of medications decays. It is known that auditory rhythmic stimulation improves gait in patients without FOG (PD-FOG), but its putative effect on patients with FOG (PD+FOG) at the end of dose has not been evaluated yet. This work evaluates the effect of auditory rhythmic stimulation on PD+FOG at the end of dose. 10 PD+FOG and 9 PD-FOG patients both at the end of dose periods, and 10 healthy controls were asked to perform several walking tasks. Tasks were performed in the presence and absence of auditory sensory stimulation. All PD+FOG suffered FOG during the task. The presence of auditory rhythmic stimulation (10% above preferred walking cadence) led PD+FOG to significantly reduce FOG. Velocity and cadence were increased, and turn time reduced in all groups. We conclude that auditory stimulation at the frequency proposed may be useful to avoid freezing episodes in PD+FOG.
The gait of people with Parkinson's disease (PD) is characterised by a number of well-defined features. From a kinematic point of view PD exhibit a reduction in step length and velocity
FOG is typical in advanced phases of the disease and it seems associated with disease duration, its grade of development, longer duration of levodopa treatment, levodopa-induced dyskinesias
Three main forms of FOG have been identified
Based on the poor correlation between FOG and UPDRS sub-scores
FOG is chiefly triggered at onset of walking and during turning, but also at narrow spaces (such as doorways) (see supporting information multimedia files
Although the effect of auditory rhythmic cueing on gait in PD is well documented
The aim of our study was to investigate the effect of rhythmic auditory stimulation on the gait of Parkinsonian patients who exhibit significant FOG (PD+FOG) during their
Participants in the study were recruited from a total of 80 patients belonging to the Asociación Parkinson Galicia and the Asociación Parkinson Ferrol (Spain). All patients were only orally medicated, without surgical operation for PD.
Patients in this group, who exhibited significant FOG (PD+FOG), had to match the following criteria:
diagnosis of idiopathic PD based on the UK Parkinson's Disease Society Brain Bank for clinical diagnostic criteria
history of freezing during walking from medical records, and score >10 (all PD displaying ≥2 in item #3) in the
predictable motor fluctuation related to dose intake, determined from medical records and examination by a neurologist
lack of auditory-visual impairment, musculoskeletal injury, and MMSE score >24
at the moment of testing, during the end of dose period, they should be able to walk 6m unaided, turn around and come back despite the freezing episodes, which should be present during preferred walking condition (un-cued, at the end of dose)
during ON periods they should be able to walk without freezing
10 volunteer PD+FOG matched the criteria and underwent the experimental protocol (6 males, 4 females; 68.20 yrs (±8.03), trochanteral height 0.89 m (±0.06), FOG
9 volunteer PD, without history of FOG (PD-FOG), were also recruited (6 males, 3 females; 64.44 yrs (±9.50), trochanteral height 0.88 m (±0.04). Inclusion criteria were the same as stated for PD+FOG, with the exception of those criteria related to FOG. The score in the FOG
10 healthy subjects (people from our institution and relatives) were selected as the Control group (8 males, 2 females; 70.20 yrs (±6.84), trochanteral height 0.89 m (±0.04); they were also screened for gait or balance impairment.
Subjects were asked to walk along a corridor (with a door in the middle), touch a button on the wall at the end, turn around, come back and touch the button on the other wall, this task in conception and distance included FOG evoking elements.
Patients came to our laboratory on two consecutive days. The first day they undertook MMSE, UPDRS
During the next day, at the end of dose, patients performed the UPDRS-III and 4 trials (2 at their preferred walking without auditory stimulation (PW) and 2 with the stimulation at a frequency 10% faster than the cadence at baseline (110A), both with the door in the middle of the corridor); healthy controls performed the 6 trials in the same day. End of dose was defined as “
The instruction given to the subjects was
All patients were evaluated in the morning after a light breakfast to avoid interference of possible protein intake at lunch, which could lead to L-dopa absorption problems. At the moment of testing PW and 110A patients confirmed to have lost the effect of medication
The recording system consisted of a series of footswitches worn as insoles in the shoes. The footswitches were connected to a radio-transmitter attached to the subjects' belt. Data (sampled at 1KHz) were sent to a receiver unit connected to the computer. This configuration allowed the stride cycle time to be registered.
Two photocells, placed 5.98 m apart, were connected to the recording system so that the records from the moment subjects crossed them were acquired. A portable in-house device provided auditory stimulation (a click) by means of headphones, which subjects wore regardless of whether or not they were stimulated. The sound was a tone with wave-frequency of 4,625 Hz, and the intensity was adjusted to be clearly perceived by the subjects without being annoying. The stimuli were delivered in pulses of 50 ms and the inter-pulse duration was customized to obtain the desired stimulation frequency.
The number of freezing episodes and their duration were measured by analysis of video footage by a specialist with 10 years experience working in a rehabilitation centre for PD, who was unaware of the protocol. Video samples were analysed by means of video software which allows frame identification (and/or sequencing) by simply keyboard strokes, allowing the identification of freezing start and end, duration and number of FOG episodes. Videos to the specialist were presented in random order and were encoded to avoid any kind of identification during evaluation; sound was off. Freezing episodes were defined following the work by Kompoliti et al.
In other to characterize FOG, the freezing episodes were grouped by duration (less than 3s; 3–10s; >10s)
Some other kinematic variables were analysed:
Velocity: Calculated as a function of the time to cover the straight section between the photocells, expressed as m/sec.
Cadence: Obtained from footswitch data corresponding to the straight section of the test, expressed as steps/sec.
Step length: Expressed in m as a function of the velocity and the cadence, again measured only over the straight section.
Turn around time: Time taken from the photocell at the end of the corridor (
The value for each kinematic variable was the mean obtained from the two trials performed in each condition.
All subjects were informed about the nature of the test and signed consent forms. The protocol was in compliance with the Helsinki declaration and was approved by the University of A Coruña Ethics Committee.
A student “t” test for independent samples was used to compare the grade of disability between the groups of patients (UPDRS-III).
One-way ANOVA was used to assess differences in motor behaviour at baseline (PW) between groups of patients and controls, also for demographics. Alternatively, a non-parametric Kruskal-Wallis test, and subsequent Mann-Whitney were performed for those variables not matching normality.
In order to determine the effect of stimulation on the kinematics, a 2x3 ANOVA model with repeated measures was performed. Two factors were defined: (i) within-subjects, (factor cue with 2 levels, PW and auditory stimulation (110A)); and (ii) between-subjects, (factor group, with 3 level PD+FOG, PD-FOG, and Controls). Given the parametric nature of this analysis a Logarithmic Transformation was performed when normality was not assumed (in the case of Turning Time for PD+FOG), so that the variables could be introduced into the analysis. Normality of distribution was assessed by means of one sample KS test.
A one-way Chi-Square (χ2) was performed in order to assess differences in proportions of type of freezing episodes. Given the task involved passing through a doorway, approaching a point, and start walking twice each trial, and just one turn, the number of three first types was adjusted by dividing each by two. Number and mean duration of the freezing episodes in the PD+FOG in presence vs. absence of stimulation were assessed by means on non-parametric Wilcoxon test. Significance was set at p≤0.05.
Differences in the UPDRS motor scores between PD+FOG and PD-FOG were not significant (t(17) = 1.163 p = 0.261); proving groups of patients were comparable in the overall disease development, (though clearly they differed in respect of the presence of FOG); demographics were not different along groups p>0.05 (age: F(2,26) = 1.305 p = 0.288; trochanteral height: F(2,26) = 0.029 p = 0.972).
During straight walking, gait patterns exhibited some other characteristic differences between groups (
PD+FOG | PD-FOG | Control | F-p values//KW | |||
0.580 (±0.313) | 0.967 (±0.214) | 1.237 (±0.160) | F(2,26) = 19.115 p<0.001 | |||
0.337 (±0.174) | 0.531 (±0.079) | 0.674 (±0.061) | F(2,26) = 20.711 p<0.001 | |||
1.727 (±0.338) | 1.819 (±0.185) | 1.831 (±0.125) | F(2,26) = 0.573 p = 0.571 | |||
26.886 (±58.690) | 3.850 (±1.083) | 2.493 (±0.557) | χ2 (2) = 18.796 p<0.001 | |||
Values: Mean, (±sd),
All the PD+FOG experienced freezing during the task. A total of 59 freezing episodes (see
(▴) At start walking; (♦) at turning; (---) at the door; (•) at approaching the target. Grey icons represent the values obtained for PD+FOG #3 and #10. The number and mean duration of the freezing episodes were significantly reduced by the presence of the stimulation when all PD were analysed (p = 0.014, and p = 0.017; respectively). When PD+FOG #3 & #10 were excluded from the analysis, in order to know if change was due to behaviour of these two extreme PD+FOG, the effect of stimulation kept on being significant, by reducing the number (p = 0.040) and mean duration (p = 0.050) of motor blocks.
PW | 110A | ||
0.927 (±0.361) | 1.008 (±0.328) | ||
0.513 (±0.182) | 0.532 (±0.154) | ||
1.792 (±0.232) | 1.878 (±0.184) | ||
11.325 (±35.212) | 4.048 (±2.500) |
PW | 110A | ||
5.900 (±6.707) |
1.400 (±1.265) |
Z = 2.446; |
|
3.119 (±4.930) |
1.020 (±1.699) |
Z = 2.395 |
PW (absence of stimulation); 110A (presence of stimulation). Values: Mean, (±sd),
Variables related to freezing are only related to PD+FOG, median and range are shown as variables were not matching normality. Stimulation led both to reduce the number and the mean duration of the freezing episodes.
The main outcome of this study is that the number of freezing episodes were significantly reduced in patients in presence of auditory stimulation, from 59 to 14 (6 when turning and 8 at start walking; Z = 2.446 p = 0.014). Mean duration was also significantly reduced (Z = 2.395 p = 0.017); see
Mean duration (a), and number (b) of the freezing episodes for each patient. PW (absence of stimulation); 110A (presence of stimulation).
It is possible that the significant reduction in the number and duration of FOG is due to change in very few subjects, rather than to the whole population. For example, PD+FOG subjects number 3 and number 10 (
For the rest of variables, stimulation proved to affect the same way all groups, as demonstrated by the lack of significant interactions cue*group. Taking this into account the stimulation led to reduce the time to turn, to increase cadence, and to increase velocity, as proved by a main effect of factor cue for each of those kinematics (see
In absence of auditory stimulation the gait of the Parkinsonian patients who “freeze” compared to those without freezing, and the latter compared to controls, exhibited lower velocity, and shorter step length, and such differences from Controls are in agreement with previous work
However, the main outcome of our study is that auditory stimulation at the frequency proposed significantly reduces the number and the mean duration of the freezing episodes in a FOG eliciting task, aimed to reproduce daily activities
The reduction in FOG is in contrast to a previous study which reported a lack of effect of auditory stimulation on FOG
The presence of the auditory stimulation interacted with the kinematic variables the same way across our different groups of subjects
Despite our results some questions about the effectiveness of cueing on FOG are still open. Here, the impact of stimulation was assessed for limited period of time, so it is pertinent to ask about its effectiveness during repeated, daily use, give the possibility of habituation to stimulation. Some work has reported rhythmic auditory stimulation entrainment in PD after a programme of auditory stimulation, modifying EMG patterns during gait
We conclude that auditory stimulation may be used in order to minimize FOG at the end of dose in affected Parkinsonian patients. Results from our study support the use of a frequency slightly above the preferred walking frequency (as measured during ON-periods in absence of FOG), which can then be used at the end of dose phase. This point strongly supports other work on the suitability of using auditory cues to improve quality of life in PD either in controlled or uncontrolled environments
Example of a patient (Example1) with motor blocks (mainly at turning) during preferred walking (no stimulation).
(7.39 MB MPG)
Example of the same patient shown in S1 (Example1) with motor blocks during auditory stimulation. Walking and turning were clearly improved.
(2.56 MB MPG)
Example of another patient (Example2) with motor blocks (mainly at crossing the door) during preferred walking (no stimulation).
(9.21 MB MPG)
Example of the same patient showed in V3 (Example2) with motor blocks (mainly at crossing the door) during auditory stimulation. Walking through the door was clearly improved.
(5.70 MB MPG)
We would like to thank Dr K.L. Grieve for his comments and corrections on the manuscript, and Dr J. Vivas for rating the motor blocks.