Speech and gait abnormalities in motor subtypes of de‐novo Parkinson's disease

Abstract Aim To investigate the presence and relationship of temporal speech and gait parameters in patients with postural instability/gait disorder (PIGD) and tremor‐dominant (TD) motor subtypes of Parkinson's disease (PD). Methods Speech samples and instrumented walkway system assessments were acquired from a total of 60 de‐novo PD patients (40 in TD and 20 in PIGD subtype) and 40 matched healthy controls. Objective acoustic vocal assessment of seven distinct speech timing dimensions was related to instrumental gait measures including velocity, cadence, and stride length. Results Compared to controls, PIGD subtype showed greater consonant timing abnormalities by prolonged voice onset time (VOT) while also shorter stride length during both normal walking and dual task, while decreased velocity and cadence only during dual task. Speaking rate was faster in PIGD than TD subtype. In PIGD subtype, prolonged VOT correlated with slower gait velocity (r = −0.56, p = 0.01) and shorter stride length (r = −0.59, p = 0.008) during normal walking, whereas relationships were also found between decreased cadence in dual task and irregular alternating motion rates (r = −0.48, p = 0.04) and prolonged pauses (r = −0.50, p = 0.03). No correlation between speech and gait was detected in TD subtype. Conclusion Our findings suggest that speech and gait rhythm disorder share similar underlying pathomechanisms specific for PIGD subtype.

impairment is a debilitating feature of PD mainly characterized by a slower velocity with shorter stride length and a compensatory increase in walking cadence. 6 These findings imply that speech timing abnormalities may have some points in common with spatiotemporal gait disturbances.
Speech and walking velocity might decrease in some PD patients, while inappropriate pauses and impaired cadence might be the result of a similar rhythmic disorder. It is thus debated to what extent speech and gait abnormalities in PD are the result of the same pathophysiology. Yet, objective data documenting relationships between speech and gait abnormalities in PD are scarce. [7][8][9][10][11] Most of the available studies have found correlations between speech dysfunction and freezing of gait in advanced PD patients on dopaminergic medication. [7][8][9] In addition, a correlation between rater-based gait festination and syllable repetition task-based oral festination has been observed. 10 The association between natural speech and walking characteristics such as velocity and rhythmicity has only been reported in PD patients treated with subthalamic nucleus deep brain stimulation. 11 However, pharmacological and surgical interventions as well as the various pace of disease progression might pose different influences on speech and gait patterns. 12,13 Potential speech and gait relations in de-novo, untreated PD patients have not yet been explored.
One possible reason for the ambiguities about the existence of similar brain processes underlying speech and gait motor functions might be the unequal proportion and severity of speech and gait disorder across individual PD patients. 6,13 These interindividual disparities in speech and gait patterns are likely related to various clinical subtypes of PD with differences in disease profiles and progression rates. 14,15 Such heterogeneity seems to be consistent with the existence of motor subtypes, most commonly divided into the postural instability/gait disorder (PIGD) or tremor dominant (TD) subtypes. 15 By definition, patients with the PIGD subtype exhibit more balance and gait impairment than patients with the TD subtype. Nevertheless, more significant speech impairment in the PIGD compared to the TD subtype has also been reported, 16 indicating that certain connections exist between speech and gait disorders in PD. Therefore, we aimed to compare temporal speech and gait parameters between the PIGD and TD motor phenotypes of de-novo PD patients and healthy controls. Additionally, we aimed to examine potential correlations between the temporal speech and gait parameters in both PD subtypes. We hypothesized that both speech and gait would be affected predominantly in the PIGD subtype and that timing abnormalities of speech and gait would be interrelated.

| Study design
From 2016 to 2022, a consecutive group of de-novo, drug-naive PD patients was recruited. PD patients were diagnosed based on the Movement Disorder Society clinical diagnostic criteria for PD 17 and investigated before the introduction of dopaminergic therapy. This study is part of a longitudinal project "biomarkers in PD (BIO-PD)" aimed to collect a large representative sample of de-novo PD patients; the detailed protocol of this project has been described previously. 18 The inclusion criteria for PD patients were as follows: (i) fulfilling criteria to allow inclusion into the TD or PIGD group, (ii) native Czech language speaker, (iii) no history of therapy with antiparkinsonian medication, (iv) no history of communication or significant neurological disorders unrelated to PD, and (v) no current or past involvement in any speech therapy. The exclusion criteria were as follows: (i) clinical or imaging signs of atypical parkinsonism and (ii) normal finding on dopamine transporter single-photon emission computed tomography examination. The healthy control subjects were recruited from the general community through advertisements. To be eligible for the study, controls had to be free of speech disorder, motor neurologic disorder, active oncologic ill-

| PD Motor subtypes classification
The PD patients were categorized into PIGD or TD clinical subtypes according to the method proposed by Stebbins et al. 21  [freezing of gait], and 3.12 [postural stability]) were used to define TD patients (ratio ≥1.15) or PIGD patients (ratio ≤0.9). In addition, patients who had a nonzero value in the numerator and a zero in the denominator were classified as TD, while patients with a zero in the numerator and a nonzero value in the denominator were classified as PIGD.

| Speech assessment
Speech recordings were performed in a quiet room with a low ambient noise level using a head-mounted condenser microphone (Beyerdynamic Opus 55, Heilbronn, Germany) placed approximately 5 cm from the subject's mouth. Speech signals were sampled at 48 kHz with 16-bit resolution. Each subject was recorded during a single session with a speech specialist. Participants were instructed to perform three vocal tasks including (i) repetition of the syllable/ pa/at least 20 times at a comfortable, self-determined, steady pace without acceleration or deceleration, (ii) fast/ta/syllable repetition for approximately 5 s per one breath, and (iii) reading a short paragraph of standardized text composed of 80 words. These vocal tasks provide comprehensive information necessary for the objective description and interpretation of motor speech disorders. 22 Each task was repeated twice per participant.

| Gait assessment
All subjects completed an extended Timed Up & Go Test (TUG): to rise from a chair, walk 10 meters at the usual preferred walking Pre-turn and post-turn gait were acquired as separate recordings on the walkway. The participants were instructed to walk under two different conditions: (i) at a normal pace (hereafter, single-task), and (ii) at a normal pace while counting down from 100 by sevens (hereafter, dual task). The rationale for including dual-task gait measure in relation to single-task speech assessment is that dual-task represents a routine way of testing gait in PD. In particular, many gait alterations in early-stage PD become apparent or exaggerated when patients are asked to walk and do another task at the same time. 6

| Acoustic speech and instrumental gait features
We performed a quantitative acoustic vocal assessment of seven distinct temporal speech dimensions ( information for the precise assessment of speech performance from the very early stages of PD. 2 Considering rhythm in syllable repetition, we examined pace acceleration using rhythm acceleration (RA) and pace instability using rhythm instability (RI). For the fast syllable repetition paradigm, we extracted slow alternating motion rates (AMR) using the diadochokinetic rate (DDKR), irregular AMR using diadochokinetic irregularity (DDKI), and imprecise consonants using voice onset time (VOT). Using reading passages, we calculated prolonged pauses using the duration of pause intervals (DPI) and abnormal articulation rate using net speech rate (NSR). The final speech values used for the statistical analyses were averaged across two repetitions to provide greater speech assessment stability. 22 Comprehensive algorithmic details on individual acoustics measures have been reported previously. 23 Also, the accuracy of algorithms for the identification of temporal intervals has been thoroughly tested in previous studies. 5,23,24 Furthermore, we selected three representative gait measures provided by GAITRite including gait velocity, cadence, and stride length ( Table 1). 6 Gait velocity was defined as a distance measured between the first footfall and the last footfall divided by the time elapsed between them, cadence as a number of steps taken within a minute provided, and stride length as the distance between the two consecutive footprints of the same foot. Gait parameters for each record (i.e., walk forward and walk back across 2 trials) were averaged. for three groups. 25

| Clinical and perceptual speech description
The TD group consisted of 40 patients (28 men The respective item of MDS-UPDRS III was equal to 1 (i.e., slight disorder); remaining subjects were scored by 0 points (i.e., without disorder). b Significantly different between TD and PIGD group with p < 0.01.

| Group differences
In the speech assessment, patients with the TD subtype showed slower AMR (DDKR: p = 0.04) as well as slower articulation rate (NSR: p = 0.03) compared to the PIGD subtype ( Figure 1A; Video S1). AMR in the TD subtype was also significantly slower (DDKR: p = 0.04) and irregular (DDKI: p = 0.03) than in controls. PIGD subtype showed greater consonant timing abnormalities (VOT: p = 0.03) compared to controls. No other significant differences between groups were found for speech assessment.
No significant differences between PIGD and TD subtypes for gait assessment were detected.

| Relationships between speech and gait
Significant correlations between speech and gait abnormalities corrected for age were found only in the PIGD subtype (Figure 2), while no relation between patterns of speech and gait disorders was observed in the TD subtype (Table 3) timing abnormalities (VOT) and gait parameters were also found for the entire PD group (Table 3).

| DISCUSS ION
In this quantitative, objective instrumental study, we attempted to detect the associations that may reflect distinctive pathophysiological mechanisms of speech and gait abnormalities in different motor subtypes of de-novo PD. In general, our findings support the assumption that objective spatiotemporal measures of gait are more affected in the PIGD subtype compared to the TD subtype.
However, in agreement with previous research, 26 this separation based on objective assessment is not fully consistent with traditional clinical classification scheme as there were large overlaps in multiple gait parameters between PIGD and TD subtypes. Considering speech assessment, current results are principally in agreement with previous research demonstrating that patients with the PIGD subtype manifest decreased voice quality, articulatory decay, and more timing abnormalities, apart from monopitch and irregular AMR presenting in the TD subtype as well. 16 Despite the effect of the TD/ PIGD phenotype was not perceptually identifiable using the MDS-UPDRS III speech item, the objective acoustic evaluation revealed that the PIGD subtype generally manifested more severe temporal speech impairment compared to the TD subtype. Likely because of more advanced and wide-ranging gait and speech disorder, the significant association between gait and speech temporal impairment across several objective measures was detected only in the PIGD subtype. The relevance of this finding can be further supported by the fact that the same strength of correlations between gait and speech parameters was not observed across the entire PD group, despite the lack of significant differences between TD and PIGD subtypes in some speech and gait variables.
Namely, in the PIGD subtype, the relationships were detected Furthermore, in the dual task in PIGD subtype, additional relationships were found between decreased cadence of gait and prolonged pauses as well as decreased diadochokinetic regularity.
These associations likely reflect rhythmic disorder that has been repeatedly reported in both gait and speech. 11,28 Similarly to the decrease of the number of steps, the PIGD patients slowed the speech tempo by increasing the length of pauses during reading and provided more variable pause gaps between following vocalizations during rapid syllable repetition. Indeed, dual-task involving serial sevens subtraction likely well represents motor-cognitive links, similarly as the production of pauses during connected speech reflects both speech-motor execution and cognitive-linguistic processing.
However, the deficits in pauses and diadochokinetic regularity in PIGD were not significantly different from the controls. Yet, these observed associations might still be explained by the fact that the dual-task walking was more cognitively challenging, as cadence in PIGD was significantly altered from controls only during the dualtask paradigm. Although we did not detect cognitive differences between PIGD and TD subtypes in our de-novo PD cohort, it is well known that PIGD subtype poses higher risk of cognitive decline over time. 29 Therefore, lower cadence in the dual task observed in our PIGD subtype might be interpreted as a subliminal precursor of future cognitive decline.
It remains to be elucidated why the observed speech and gait associations occurred in the PIGD subtype but not in the TD subtype. Patients with PIGD tend to exhibit a more rapid decline in motor and cognitive function compared to TD, which is considered   30 Neuroanatomically, distributed regional gray matter atrophy has been demonstrated in the PIGD, with the disruption in the precentral gyrus hypothesized to be involved in postural control, balance, and gait. 31 In this context, it is noteworthy that acquired motor speech impairment has traditionally been associated with damage to contiguous regions, namely the left inferior precentral cortex, as corroborated by modern imaging studies. 32 Speech disorder thus might be attributed to different pathophysiological mechanisms underlying the appendicular (limb tremor, bradykinesia, and rigidity) and axial motor symptoms of PD. In particular, temporal speech deviations such as prolonged voicing onset time seem to represent the axial speech symptom as it does not respond to dopaminergic medication and thus cannot be solely attributable to more advanced nigrostriatal degeneration. 13 Therefore, as in the case of gait dysfunction, more widespread brain damage affecting extranigral cortical and/or subcortical regions appears to contribute to temporal speech deviations. We may thus hypothesize that more advanced alpha-synuclein pathology spreading into cortical areas, consistent with Braak stage 4 or higher, contributes to shared pathomechanisms underlying temporal speech and gait dysfunction.
Previous findings have been mostly limited to observed correlations between freezing of gait and freezing of speech (i.e., stutteringlike speech dysfluencies) or oral festination and festination of gait in later PD stages. 7-10 The present early-stage PD cohort was free of such abnormalities of gait and speech during clinical examination, and particularly did not show any tendency toward speech acceleration as measured via the rhythmic speech paradigm. This agrees with previous studies reporting low occurrence of oral festination events in early-stage PD, 33 as well as the correlation of the severity of pace acceleration to motor deficits and disease duration. 24,28 Accordingly, it has been shown that PD patients exhibited significantly more dysfluent speech events compared to controls after an average time of 4.5 years of dopaminergic therapy after diagnosis. 34 Oral festination and short "rushes" of speech in PD are closely connected with changes in speaking rate. 35 Our observation of F I G U R E 2 Significant correlations between speech and gait measures in the PIGD subtype. The blue circles demonstrate the real, uncorrected gait and speech values while the correlation coefficient r and its corresponding p-value are corrected to age. DDKI, diadochokinetic irregularity; DPI, duration of pause intervals; VOT, voice onset time. ing rate in PD in the literature. 36 This finding appears to be well following differences between PD motor subtypes detectable on neuroimaging. Structural connectivity analysis has shown that patients with PIGD have alterations of cortico-basal ganglia pathways while TD do not. 37 Therefore, more cortical brain atrophy is likely responsible for the faster speech rate observed in the PIGD subtype, which might be a precursor of events such as oral festination that may develop later in the disease process. On the other hand, TD patients show greater cerebello-thalamo-cortical circuitry dysfunction. 38,39 Previous study has demonstrated that the extent of cerebellar atrophy is related to a reduced rate, 40 likely due to a reduced ability to program movement sequences in advance of movement onset. 41 These assumptions can be further supported by neuroimaging finding that the various cortical and subcortical brain regions engaged in speech motor control might be organized into two separate networks (medial and dorsolateral premotor cortex, anterior insula, and superior cerebellum versus sensorimotor cortex, basal ganglia, and inferior cerebellum), most presumably responsible for motor preparation and execution processes. 42 Since both loops are involved in the rate of the produced syllable trains and associated rhythmic modulations, 42 they might also help to explain clinical observations of accelerated or even slowed speaking rate in PD.
However, it is challenging to differentiate the problems from rhythm generation (internal cueing of rhythm) and the initiation/bradykinesia issues that arise from external/motoric factors.
A limitation of this single-center study is that our sample comprises only 20 patients with PIGD, as this subtype is relatively rare in de-novo PD. This is in accordance with the PPMI cohort consisting of a large sample of de-novo PD, where TD represents the dominant TA B L E 3 Correlations between speech and gait measures in the TD and PIGD subtypes and entire PD group.  subtype compared to PIGD. 43 Thus, future research based on a greater sample, preferably obtained via multicenter cohorts, should confirm and further elaborate our findings. It would also be preferable to support the present findings by neuroimaging data to substantiate the cortical involvement in gait/speech rhythm disorder.
In conclusion, this study underlines several common features of gait and speech abnormalities in PD specific to the PIGD sub-

CO N FLI C T O F I NTE R E S T S TATE M E NT
All authors report no conflict of Interest concerning the research related to the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Individual participant data that underlie the findings of this study are available upon request to the corresponding author by qualified researchers (i.e., affiliated to a respected university or research institution/hospital). The speech data are not publicly available due to their contain of information that could compromise the privacy of study participants.