Manual dexterity is an important component in a comprehensive assessment of upper extremity function in both healthy subjects and during rehabilitation of patients suffering from chronic ischemic stroke and Parkinson’s (PD) patients (Grice, et al. 2003). The assessment of the upper limb’s movements and the improvement’s measurement of the subject during the period of training can be done accurately and legitimately if validated protocols are used. Nowadays, for manual dexterity there are several tests that have been already validated. Nevertheless, all these tests have mostly a single outcome — for example a measure of time — that not only it may not be enough sensitive to characterize manual dexterity in healthy subjects, but also it may not be a repeatable and reliable measure. Moreover, these tests are mostly applied to mobility assessment merely in terms of kinematic output, but they barely analyze the aspects concerning motor control and movement’s programming. For this reason, the integration of kinematics data obtained with optoelectronic system and an ulterior electroencephalographic (EEG) recording during the task may add new outcome variables to make completer and more comprehensive the results. The Nine Hole Peg Test (NHPT) is widely considered a gold standard metric for manual dexterity (Feys, et al. 2017). Performance of the standard NHPT requires the pegs to be picked up from the container one-by-one unimanually, transferred and inserted into the holes of the pegboard until it is filled. The test is performed as quickly as possible and the only outcome variable is the total time complete to the task, measured with a stopwatch by the examiner. Nowadays, there are not validated protocols that assess kinematic measurements during the execution of the commercially available NHPT or especially that combine kinematic measurements and cortical activity recording during the same task. Compared to the standard NHPT, this innovative instrumental version of the NHPT allows to obtain a more detailed and sensitive assessment of manual dexterity in healthy subjects. Moreover, this may be a first step towards the application of the proposed protocols on patients with neurological disorders or suffering from chronic ischemic stroke. For these reasons, the main aim of this study is to develop and validate a kinematic protocol to assess manual dexterity in a large group of healthy subjects. The proposed protocol allows the investigation of upper limb’s kinematics during the execution of the instrumental version of the commercially available NHPT through the use of optoelectronic system and these specifics make it suitable to the future integrative measure with EEG. Twenty-five right-handed, young and healthy adults were enrolled to participate to this study. The whole task consisted in the execution of the test firstly twice with the right hand and then twice with the left hand, as quickly as possible. For each subject, two sessions of the task were performed, one in the morning and the other one in the afternoon on the same day, with about 6 hours of break, resulting in four trials per session — twice per hand — and eight acquisition per subject in total. All the acquisitions have been realized at the laboratory “Luigi Divieti” of the Department of Bioengineering of Politecnico di Milano (Milan, Italy). For the kinematic acquisition thirteen passive retroreflective markers were affixed on specific landmarks of the subject: left and right acromial angles; jugular notch; left and right lateral epicondyle; left and right ulnar styloid; left and right distal epiphysis of the radius; between the left and right heads of the second and third metacarpals; left and right distal phalanx. Kinematic analysis was recorded using an 8-camera motion capture system (SMART DX-400, BTS, Italy) and data analysis was conducted using the software SMART Analyzer (BTS, Italy). The kinematics during peg insertion was described in four different test sub-phases: Peg Grip, Peg Transfer, Peg in Hole and Hand Return. After the last peg is inserted the Peg Removing phase starts. The beginning/end of each phase was defined manually from the track of the index finger marker in the Z-direction. Twenty-three kinematic outcomes were extracted from each acquisition: test total duration, Peg Removing phase duration, Peg Grip phase total time, Peg Grip phase mean time, Peg Transfer phase total time, Peg Transfer phase mean time, Peg in Hole phase total time, Peg in Hole phase mean time, Hand Return phase total time, Hand Return phase mean time, jerk during Peg Transfer phase and Hand Return phase, mean velocity during Peg Transfer phase and Hand Return phase, velocity peak during Peg Transfer phase and Hand Return phase, time to peak during Peg Transfer phase and Hand Return phase, range of motion (ROM) of the trunk along all the three direction, index of curvature during Peg Transfer phase and Hand Return phase. Four statistical analyses were carried out: test-retest reliability, between all the first trails with both hands and all the second trials with both hands; intra-rater reliability, between the best performances with both hands of the first session and the best performances, with respect to the total time of the execution, with both hands in the second session; convergent validity, in particular the correlation between each kinematic outcome and the standard NHPT outcome, i.e. test total time; discriminant validity, between the best performances with the dominant hand and the best performances with the non-dominant hand. For test-retest and intra-rater reliability, the Intraclass Correlation Coefficient (ICC) — ICC 2,k for test-retest reliability and ICC 3,k for intra-rater reliability —, the 95% confidence interval, the Standard Error of Measurement (SEM) and the %SEM were calculated. Convergent validity was instead evaluated deriving the Pearson’s correlation coefficient or Spearman’s correlation coefficient, depending on data distribution. Comparisons between dominant and non-dominant trials were made with paired t tests or the Wilcoxon test, depending on the data distribution. Results from the test-retest reliability show an excellent, good or at least moderate ICC for all the kinematic parameters, indicating that the proposed protocol is consistent within the same session from one test administration to the next, where a single rater uses the same methods and the same testing conditions. This is a huge result and it indicates that adding new outcomes to the test leads to an improvement in the test-retest reliability in healthy subjects, in contrast with the low to moderate reliability of the standard NHPT (Mathiowetz, et al. 1985) (Grice, et al. 2003). Moreover, the statistical analysis results also an excellent, good or moderate intra-rater reliability for all kinematic outcomes, with the exception of the ROM of the trunk along X that shows a poor ICC. This result highlights that this instrumental version of the NHPT, despite the much more additional outcome variables with the respect to the standard NHPT, is however reliable and repeatable also between sessions, in response to the repositioning of the retroreflective markers by the same operator. The convergent validity analysis shows for most of the kinematic outcomes a very strong, moderate or fair correlation with the test total time. Only for a few parameters — time to peak, trunk ROMs and Indexes of Curvature — the correlation is poor, as a result of a limited variability of these kinematic parameters in healthy subjects, in agreement with literature (Johansson and Häger 2019). Lastly, from statistical analysis results, we can claim that this proposed protocol is able to discriminate significantly from dominant and non-dominant hand in most of its outcomes. This means that adding more variables than the standard one provides a better and more deeply characterization of subject’s manual dexterity. However, the test is not able to discriminate between right hand and left hand for mean velocity, trunk ROMs along X and Z and index of curvature outcomes, because of the limited variability in these parameters between the two hands in healthy subjects. Proving the good validity and reliability of this instrumental version of the NHPT allows to future applications in pathological subjects, where the increased variability of the outcome variables may enhance even more the importance to have much more valid and reliable kinematic parameters than the single outcome of the standard NHPT. Furthermore, this kinematic protocol is the first step towards a more complete and remarkable protocol, in which both kinematics and cortical activity are recorded simultaneously during the NHPT. Regarding this last protocol only a pilot study with few subjects was carried out in this dissertation, in order to briefly enlighten the pros and limitations of EEG integration with kinematics recording. The NHPT was completed in the exact same way of the first protocol, but the subject performed only one session (twice execution with right hand and twice with left hand). The only difference was the introduction of the EEG recording — using a brain cap with 61 channels — combined with optoelectronic system. In addition, in this pilot study, it was also explored cortical activity during action observation (AO) and motor imagery (MI) tasks, in order to study motor neuron activation during the observation or imagination of simple motor tasks of manual dexterity. Indeed, this work is part of a larger study that aims to demonstrate the effectiveness of action observation therapy (AOT) followed immediately by sleep, in order to improve and enhance manual dexterity in healthy subjects. Lately, there have been actually studies supporting and validating the efficacy of AO regarding both the improvement of a specific motor skill in healthy subjects and the rehabilitation of patients suffering from chronic ischemic stroke, PD patients, children with cerebral palsy, and non-neurological patients such as those undergoing orthopedic surgery of the hip or knee (Brunsdon, Bradford and Ferguson 2019) (Gatti, et al. 2013) (Zhu, et al. 2015). Concurrently, MI has been studied widely along with AO and it is also considered an effective tool (Tani, et al. 2017) (Marusic, et al. 2018). EEG pre-processing and data analysis were conducted with the collaboration of the research group of B3Lab – Biosignals, Bioimaging, Bioinformatics of the Politecnico di Milano. In particular, data’s noisy and irrelevant segments were manually eliminated and Independent Component Analysis (ICA) was applied to separate brain activity and noise sources. In order to determine the subject’s α frequency range, it was examined EEG recording during the eyes closed state when α rhythm is dominant, while the individual μ-rhythms needed the analysis of the open eyes’ baseline over the central channels that correspond to the motor area. Then, during the NHPT execution the relative power in the classical EEG frequency bands and the μ -band were extracted. Event-Related Spectral Perturbation (ERSP) maps were instead used to identify subject-specific frequency bands reactive to action observation and motor imagery (Tacchino, et al. 2016). Results from EEG recording during the manual motor task show that the μ rhythm, even if it is similar to the α rhythms, has two particularities: in the power maps during rest with eyes open the μ rhythm is more localized on the central channels with respect to the α power which spread posteriorly; moreover, a reduction of central μ power is clearly visible during movement. However, in this pilot study, we’ve encountered also some limitations in the integration of EEG system with the optoelectronic system. Some of them are purely technical problems regarding the experimental setup made tricky by the unwieldy of the EEG system that may interfere with subject’s movement or markers’ visibility, resulting sometimes in artifacts in the EEG signal. Others drawbacks concerns the fact that the NHPT is a simple and fast motor task and for this reason few outcome parameters can be derived from cortical activity. Nevertheless, some of these limitations may be easily overcome and this integration of EEG in kinematics recording can be a starting point and a solid basis for future investigation of the assessment of manual dexterity in both healthy and pathological subjects. On the other hand, from EEG recording during AO and MI tasks, we’ve found that during AO the desynchronization in the μ and β ranges is evident — even if the significance was not tested here —, whereas during MI this event is less observable. This distinction can be mostly observed in the ERSP images of C4, CP3 and CP4 electrodes, where the desynchronization is much more present during AO than MI. An explanation of that could be that during AO also visuo-motor system is involved, whereas in MI there isn’t a precise instant we can use to synchronize the EEG epoch since the subject can start imaging whenever she/he wants after the cue. This brief experimental trial has allowed us to point out the effectiveness of AO and MI of a simple motor task, since we can observe neuron activity also during observation or imagination. Furthermore, from this short study on AO and MI we’ve remarked that these innovative therapies can easily and efficiently be included in a period of training for healthy subjects to improve manual dexterity or rehabilitation for pathological patients to recover upper arm’s ability. In conclusion, in this dissertation we’ve carried out and demonstrate that the proposed instrumental version of the NHPT with kinematic recording has an acceptable test-rest reliability, intra-rater reliability, convergent validity and discriminant validity. For this reason, this protocol can be included in further applications in order to thoroughly assess manual dexterity in healthy subjects. Moreover, the same protocol may be in the future applicated and validated also with pathological patients. Moreover, in this work we’ve also mentioned and proposed an ulterior version of the instrumented NHPT, where not only kinematics but also cortical activity has been recorded. Despite some minor limitations, this innovative protocol may be a solid basis for future applications and a deeper understanding of manual dexterity both in healthy and pathological subjects. Lastly, we’ve also investigated cortical activity during AO and MI, reaching results in accordance with literature. All these studies are a starting point and a solid basis for a larger study that aims to demonstrate the effectiveness of AOT and MI followed immediately by sleep, in order to improve and enhance manual dexterity in healthy subjects.
La destrezza manuale è un importante elemento per una esaustiva valutazione della funzione motoria dell’arto superiore sia in soggetti sani sia nell’ambito di una terapia di riabilitazione di pazienti affetti da ictus ischemico cronico o malattia di Parkinson (Grice, et al. 2003). La valutazione della mobilità dell’arto superiore e del suo miglioramento a seguito di un periodo di allenamento può essere eseguita accuratamente e legittimamente solo se i protocolli applicati sono validati. Al giorno d’oggi sono già stati validati diversi test per la valutazione della destrezza manuale. Tuttavia, la maggior parte di questi ha una singola misura di uscita, per esempio una misura di tempo, che potrebbe sia non essere abbastanza sensibile da caratterizzare la destrezza manuale in soggetti sani sia non essere sufficientemente ripetibile ed affidabile. Spesso questi test, inoltre, si rivolgono alla sola valutazione del movimento in termini di output cinematico, ma scarsamente analizzano gli aspetti più legati al controllo motorio e alla programmazione del movimento. Per questo motivo l’integrazione di dati cinematici ottenuti con il sistema optoelettrico e un ulteriore rilevamento del segnale elettroencefalografico (EEG) durante il test motorio potrebbero aggiungere nuove grandezze per una più completa ed esaustiva la valutazione finale. Il Nine Hole Peg Test (NHPT) è comunemente considerato il “gold standard” per la valutazione della destrezza manuale (Feys, et al. 2017). L’esecuzione del test consiste nel raccogliere tutti e nove i pioli con una sola mano uno ad uno dal contenitore apposito e inserirli nei nove buchi. Il test deve essere eseguito il più velocemente possibile e l’unica misura di uscita è il tempo totale necessario al soggetto per concludere il test, misurato con un cronometro dall’esaminatore. Fino ad oggi non sono stati ancora validati protocolli che includono l’acquisizione di misure cinematiche durante l’esecuzione del NHPT commerciale e a maggior ragione che combinano l’acquisizione della cinematica e quella dell’attività corticale contemporaneamente durante il test. In confronto alla tradizionale esecuzione del NHPT, questa innovativa versione strumentale permette di ottenere una valutazione della destrezza in soggetti sani estremamente dettagliata e sensibile. Inoltre, questo protocollo potrebbe considerarsi un primo passo verso una futura applicazione di esso con pazienti affetti da malattie neurodegenerative o da ictus ischemico cronico. Per i motivi sopra citati, il principale obiettivo di questo studio è di implementare e validare un protocollo cinematico con lo scopo di valutare la destrezza manuale in un ampio gruppo di soggetti sani. Il protocollo proposto permette uno studio dettagliato della cinematica dell’arto superiore durante l’esecuzione della versione strumentale del NHPT commerciale tramite l’utilizzo di un sistema optoelettrico; queste specifiche lo rendono particolarmente adeguato alla futura misura integrata dell’EEG. In particolare, per questo studio sono stati reclutati 25 soggetti giovani, sani e destrimani. Una sessione del test consiste inizialmente nell’esecuzione del NHPT due volte con la mano destra e successivamente due volte con la mano sinistra, sempre il più velocemente possibile. Ciascun soggetto è quindi sottoposto a due sessioni, una in mattinata e una nel pomeriggio dello stesso giorno, con una pausa intermedia di circa sei ore; al termine delle due sessioni si ottengono quindi otto acquisizioni per soggetto, quattro per sessione. Tutte le acquisizioni sono state realizzate al laboratorio “Luigi Divieti” del Dipartimento di Bioingegneria del Politecnico di Milano (Italia). Per l’acquisizione della cinematica sono stati applicati in totale tredici marker passivi catarifrangenti su specifiche posizioni anatomiche del soggetto: angoli acromiali destro e sinistro, incisura giugulare, epicondilo laterale destro e sinistro, stiloide dell’ulna destra e sinistra, epifisi distale del radio destro e sinistro, tra gli estremi destro e sinistro del secondo e terzo metacarpo, falange distale destra e sinistra. L’analisi cinematica è stata registrata utilizzando un sistema a otto telecamere (SMART DX-400, BTS, Italia) e l’analisi dei dati è stata seguita con il software SMART Analyzer (BTS, Italia). La cinematica durante l’inserimento dei pioli nei buchi è stata suddivisa in quattro sottofasi: presa del piolo, trasferimento del piolo, inserimento del piolo nel buco e ritorno della mano. Dopo l’inserimento del nono e ultimo piolo ha inizio quindi la fase di rimozione dei pioli. Nel protocollo cinematico l’inizio e la fine di ciascuna fase sono stati definiti manualmente sulla traccia del marker dell’indice della mano testata nella direzione Z. Ventitré variabili cinematiche sono state estratte da ciascuna acquisizione: durata totale del test, durata della fase di rimozione dei pioli, tempo totale delle fasi di presa del piolo, tempo medio delle fasi di presa del piolo, tempo totale delle fasi di trasferimento del piolo, tempo medio delle fasi di trasferimento del piolo, tempo totale delle fasi di inserimento del piolo, tempo medio delle fasi di inserimento del piolo, tempo totale delle fasi di ritorno, tempo medio delle fasi di ritorno, jerk durante le fasi di trasferimento e ritorno, velocità media dell’indice durante le fasi di trasferimento e ritorno, picco di velocità dell’indice durante le fasi di trasferimento e ritorno, time to peak (tempo necessario a raggiungere il picco di velocità, espresso in percentuale rispetto alla durata totale della fase considerata) durante le fasi di trasferimento e ritorno, ROM (range of motion, ampiezza di movimento) del tronco lungo le tre direzioni, indice di curvatura durante le fasi di trasferimento e ritorno. Sono stati quindi eseguiti quattro tipi di analisi statistica: affidabilità test-retest, mettendo a confronto tutte le prime prove con entrambe le mani e tutte le seconde prove con entrambe le mani; affidabilità intra-rater, confrontando tutte le miglior prove (considerando il tempo totale dell’esecuzione del test) con entrambe le mani della prima sessione e tutte le miglior prove con entrambe le mani della seconda sessione; validità convergente, cioè la correlazione tra ciascuna variabile cinematica e la tradizionale misura di uscita del NHPT, la durata totale del test; validità discriminante, confrontando le miglior prove con la mano dominante e le miglior prove con la mano non dominante. Per valutare l’affidabilità test-retest e l’affidabilità intra-rater, sono stati calcolati il coefficiente di correlazione intraclasse (ICC di tipo 2,k nell’analisi di affidabilità test-retest e di tipo 3,k nell’analisi di affidabilità intra-rater), l’intervallo di confidenza al 95%, l’errore standard di misurazione (SEM) e l’errore standard di misurazione in percentuale (%SEM). La validità convergente è stata invece valutata attraverso il coefficiente di correlazione di Pearson o il coefficiente di correlazione di Spearman, a seconda della distribuzione dei dati. Il confronto tra le prove effettuate con la mano destra e quelle con la mano sinistra è stato portato a termine applicando il t test a campioni appaiati o il test di Wilcoxon, a seconda della distribuzione dei dati. I risultati dell’analisi di affidabilità test-retest mostrano un eccellente, buono o almeno moderato coefficiente di correlazione intraclasse per tutti i parametri cinematici, denotando che il protocollo proposto è consistente e ripetibile tra le prove della stessa sessione, in cui un solo operatore utilizza gli stesso metodi e le stesse condizioni sperimentali. Quest’ultimo è un risultato soddisfacente e dimostra come, aggiungendo nuove variabili al test, si hanno miglioramenti nell’affidabilità test-retest in soggetti sani, in contrasto con la bassa o moderata affidabilità riscontrata nel tradizionale NHPT (Mathiowetz, et al. 1985) (Grice, et al. 2003). Inoltre, dall’analisi statistica risulta anche una eccellente, buona o moderata affidabilità intra-rater per tutti i parametri cinematici, con l’eccezione dello scarso ICC risultante dal ROM del tronco lungo X. Questo risultato evidenzia come la versione strumentale del NHPT, nonostante l’aggiunta di molte più variabili di uscita rispetto al test tradizionale, è comunque affidabile e ripetibile anche tra diverse sessioni a seguito di un riposizionamento dei marker dallo stesso operatore. L’analisi di validità convergente rileva per la maggior parte dei parametri una correlazione molto robusta, moderata o discreta con la durata totale del test. Solo per pochi parametri (time to peak, ROMs del tronco e indici di curvatura) la correlazione risulta scarsa, poiché, in accordo anche con la letteratura (Johansson and Häger 2019), la loro variabilità è molto limitata in soggetti sani. Per ultimo, dai risultati dell’analisi statistica, si può affermare che il protocollo proposto è anche in grado di discriminare significativamente la mano dominante dalla mano non dominante in quasi tutti i parametri cinematici. Questo significa che l’aggiunta di più variabili rispetto al test tradizionale fornisce una caratterizzazione migliore e più esaustiva della destrezza manuale del soggetto. Il test non è però in grado di discriminare destra e sinistra per alcuni parametri (ROMs del tronco lungo X e Z e indice di curvatura), a causa sempre della limitata variabilità riscontrata in questi valori durante le prove con entrambe le mani in soggetti sani. In conclusione, la validazione della versione strumentale del NHPT consente eventuali applicazioni future del protocollo anche con soggetti patologici, dove una maggior variabilità dei parametri cinematici in esame potrebbe migliorare ancora di più l’importanza di avere una vasta gamma di parametri cinematici validi e affidabili, rispetto alla singola misura di tempo del NHPT tradizionale. Inoltre, il protocollo cinematico proposto è il primo passo verso un protocollo più completo che integra contemporaneamente l’acquisizione cinematica e quella dell’attività corticale. In questo lavoro, il protocollo comprendente sia la cinematica sia l’EEG è stato applicato solo ad uno studio pilota con pochi soggetti, con lo scopo di discutere brevemente i vantaggi e svantaggi dell’integrazione dell’EEG alla cinematica. In questo protocollo, il test NHPT deve essere completato dal soggetto con le stesse modalità descritte precedentemente, ma viene eseguito solo in una sessione (due volte con la mano destra e due volte con la mano sinistra). L’unica differenza è quindi l’introduzione della registrazione del segnale EEG utilizzando una cuffia a 61 canali. Nello studio pilota, oltre all’esecuzione del test, è stata anche studiata l’attività corticale durante esercizi di Action Observation (AO) e Motor Imagery (MI), con lo scopo di osservare l’attività neuronale durante l’osservazione o l’immaginazione di semplici esercizi motori di destrezza manuale. Infatti, questo lavoro fa parte di un più ampio studio che mira a dimostrare l’efficacia della terapia di AO seguita immediatamente dal sonno con l’obiettivo di migliorare e rafforzare la destrezza manuale in soggetti sani. Negli ultimi anni, infatti, sono stati pubblicati molti studi che supportano e validano l’efficacia dell’AO sia nel miglioramento di una specifica capacità motoria in soggetti sani sia nella riabilitazione di pazienti affetti da ictus ischemico cronico, malattia di Parkinson, bambini con paralisi cerebrale o patologie non neurologiche come pazienti sottoposti a interventi chirurgici ortopedici al ginocchio o all’anca (Brunsdon, Bradford and Ferguson 2019) (Gatti, et al. 2013) (Zhu, et al. 2015). Allo stesso tempo, insieme all’AO, anche la MI è stata ampliamente studiata e considerata uno strumento efficace (Tani, et al. 2017) (Marusic, et al. 2018). L’elaborazione e l’analisi del segnale EEG è stata eseguita in collaborazione con il gruppo di ricerca del B3Lab (Biosignals, Bioimaging, Bioinformatics) del Politecnico di Milano. In particolare, le parti contenenti rumore o comunque irrilevanti per il nostro scopo sono state manualmente eliminate; è stata inoltre applicata l’analisi delle componenti indipendenti (ICA) con l’obiettivo di separare l’attività corticale dalle sorgenti di rumore. Per determinare il range di frequenza di tipo α del soggetto è stata presa in esame la registrazione dell’EEG durante la condizione di riposo ad occhi chiusi, in cui il ritmo α è dominante; per l’individuazione dei ritmi μ, invece, è stata necessaria l’analisi del segnale ad occhi aperti sui canali centrali che corrispondono all’aerea motoria. Successivamente, durante l’esecuzione del NHPT, è stata estratta la potenza relativa nelle bande di frequenza classiche e nella banda μ. Le mappe di Event-Related Spectral Perturbation (ERSP) sono state invece utilizzate per identificare le bande specifiche del soggetto durante gli esercizi di AO e MI (Tacchino, et al. 2016). I risultati dell’analisi EEG durante il NHPT mostrano che il ritmo μ, anche se simile al ritmo α, ha due particolarità: nelle mappe di potenza durante la condizione di riposo ad occhi aperti il ritmo μ è più localizzato sui canali centrali rispetto al ritmo α che si espande posteriormente; inoltre, una riduzione della potenza centrale μ è chiaramente visibile durante il movimento. In ogni caso, nello studio pilota in esame, si sono incontrati anche alcuni ostacoli nell’integrazione del sistema EEG con quello optoelettronico. Alcuni di questi sono problemi di natura puramente tecnica riguardanti il setup sperimentale reso difficoltoso dall’ingombro del sistema EEG che in alcuni casi ha rischiato di interferire con il movimento del soggetto o la visibilità dei marker causando artefatti nel segnale in uscita. Altre limitazioni riguardano il fatto che il NHPT è un test motorio semplice e veloce e per questo dall’attività corticale possono essere rilevati solo pochi parametri. Tuttavia, alcune di queste limitazioni possono essere facilmente superate e l’integrazione del sistema EEG con quello cinematico può essere un punto di inizio e una solida base per studi futuri sulla destrezza manuale in soggetti sani o patologici. Dal segnale EEG registrato durante AO and MI, invece, si è visto che durante l’osservazione di un’azione motoria vi è un’evidente desincronizzazione nei range μ e β, mentre durante l’immaginazione della stessa azione questo evento è meno visibile. Questa distinzione può essere maggiormente osservata nelle mappe ERSP degli elettrodi C4, CP3 e CP4, dove la desincronizzazione è molto più presente durante AO rispetto a MI. Una possibile spiegazione potrebbe essere che durante l’osservazione dell’azione è coinvolto anche il sistema visivo-motore, mentre durante l’immaginazione non c’è un istante preciso che si può usare per sincronizzare il segnale EEG in quanto il soggetto può iniziare ad immaginare quando vuole dopo il segnale apposito. Questo lavoro sperimentale ci ha permesso di notare l’efficacia dell’AO and MI di un semplice esercizio motorio, in quanto si è potuto osservare l’attività neuronale sia durante l’osservazione sia l’immaginazione. Inoltre, da questo breve studio si è rimarcato come queste terapie innovative possano facilmente ed efficientemente essere incluse sia in un periodo di allenamento per soggetti sani con lo scopo di migliorare la destrezza manuale, sia in un periodo di riabilitazione per pazienti patologici con lo scopo di recuperare la mobilità dell’arto superiore. In conclusione, in questa tesi si è dimostrato che la versione strumentale del NHPT con la sola acquisizione cinematica ha affidabilità test-retest, affidabilità intra-rater, validità convergente e validità discriminante accettabili. Per questo, il protocollo proposto può essere incluso in future applicazioni con lo scopo di valutare esaustivamente la destrezza manuale in soggetti sani. Inoltre, lo stesso protocollo potrebbe essere applicato e validato in futuro anche con soggetti patologici. In questa tesi, si è anche proposta un’ulteriore versione del NHPT in cui viene registrata non solo la cinematica ma anche l’attività corticale del soggetto. Nonostante alcune piccole limitazioni, questo protocollo innovativo può diventare un fondamento per future applicazioni e uno strumento per raggiungere una ancora più esaustiva conoscenza della destrezza manuale sia in soggetti sani sia in pazienti patologici. Infine, in questa tesi, si è anche osservata l’attività corticale durante AO e MI, raggiungendo risultati e conclusioni conformi alla letteratura. Tutti gli studi portati avanti in questo lavoro mirano ad essere quindi un punto di partenza ed una solida base per l’implementazione di un più ampio studio che vuole dimostrare l’efficacia della terapia di AO e MI seguita immediatamente dal sonno con lo scopo di migliorare la destrezza manuale in soggetti sani.
Validation of the instrumental version of nine hole peg test in healthy subjects : a starting point for EEG integration
Monfardini, Erica;Lozza, Ilaria
2019/2020
Abstract
Manual dexterity is an important component in a comprehensive assessment of upper extremity function in both healthy subjects and during rehabilitation of patients suffering from chronic ischemic stroke and Parkinson’s (PD) patients (Grice, et al. 2003). The assessment of the upper limb’s movements and the improvement’s measurement of the subject during the period of training can be done accurately and legitimately if validated protocols are used. Nowadays, for manual dexterity there are several tests that have been already validated. Nevertheless, all these tests have mostly a single outcome — for example a measure of time — that not only it may not be enough sensitive to characterize manual dexterity in healthy subjects, but also it may not be a repeatable and reliable measure. Moreover, these tests are mostly applied to mobility assessment merely in terms of kinematic output, but they barely analyze the aspects concerning motor control and movement’s programming. For this reason, the integration of kinematics data obtained with optoelectronic system and an ulterior electroencephalographic (EEG) recording during the task may add new outcome variables to make completer and more comprehensive the results. The Nine Hole Peg Test (NHPT) is widely considered a gold standard metric for manual dexterity (Feys, et al. 2017). Performance of the standard NHPT requires the pegs to be picked up from the container one-by-one unimanually, transferred and inserted into the holes of the pegboard until it is filled. The test is performed as quickly as possible and the only outcome variable is the total time complete to the task, measured with a stopwatch by the examiner. Nowadays, there are not validated protocols that assess kinematic measurements during the execution of the commercially available NHPT or especially that combine kinematic measurements and cortical activity recording during the same task. Compared to the standard NHPT, this innovative instrumental version of the NHPT allows to obtain a more detailed and sensitive assessment of manual dexterity in healthy subjects. Moreover, this may be a first step towards the application of the proposed protocols on patients with neurological disorders or suffering from chronic ischemic stroke. For these reasons, the main aim of this study is to develop and validate a kinematic protocol to assess manual dexterity in a large group of healthy subjects. The proposed protocol allows the investigation of upper limb’s kinematics during the execution of the instrumental version of the commercially available NHPT through the use of optoelectronic system and these specifics make it suitable to the future integrative measure with EEG. Twenty-five right-handed, young and healthy adults were enrolled to participate to this study. The whole task consisted in the execution of the test firstly twice with the right hand and then twice with the left hand, as quickly as possible. For each subject, two sessions of the task were performed, one in the morning and the other one in the afternoon on the same day, with about 6 hours of break, resulting in four trials per session — twice per hand — and eight acquisition per subject in total. All the acquisitions have been realized at the laboratory “Luigi Divieti” of the Department of Bioengineering of Politecnico di Milano (Milan, Italy). For the kinematic acquisition thirteen passive retroreflective markers were affixed on specific landmarks of the subject: left and right acromial angles; jugular notch; left and right lateral epicondyle; left and right ulnar styloid; left and right distal epiphysis of the radius; between the left and right heads of the second and third metacarpals; left and right distal phalanx. Kinematic analysis was recorded using an 8-camera motion capture system (SMART DX-400, BTS, Italy) and data analysis was conducted using the software SMART Analyzer (BTS, Italy). The kinematics during peg insertion was described in four different test sub-phases: Peg Grip, Peg Transfer, Peg in Hole and Hand Return. After the last peg is inserted the Peg Removing phase starts. The beginning/end of each phase was defined manually from the track of the index finger marker in the Z-direction. Twenty-three kinematic outcomes were extracted from each acquisition: test total duration, Peg Removing phase duration, Peg Grip phase total time, Peg Grip phase mean time, Peg Transfer phase total time, Peg Transfer phase mean time, Peg in Hole phase total time, Peg in Hole phase mean time, Hand Return phase total time, Hand Return phase mean time, jerk during Peg Transfer phase and Hand Return phase, mean velocity during Peg Transfer phase and Hand Return phase, velocity peak during Peg Transfer phase and Hand Return phase, time to peak during Peg Transfer phase and Hand Return phase, range of motion (ROM) of the trunk along all the three direction, index of curvature during Peg Transfer phase and Hand Return phase. Four statistical analyses were carried out: test-retest reliability, between all the first trails with both hands and all the second trials with both hands; intra-rater reliability, between the best performances with both hands of the first session and the best performances, with respect to the total time of the execution, with both hands in the second session; convergent validity, in particular the correlation between each kinematic outcome and the standard NHPT outcome, i.e. test total time; discriminant validity, between the best performances with the dominant hand and the best performances with the non-dominant hand. For test-retest and intra-rater reliability, the Intraclass Correlation Coefficient (ICC) — ICC 2,k for test-retest reliability and ICC 3,k for intra-rater reliability —, the 95% confidence interval, the Standard Error of Measurement (SEM) and the %SEM were calculated. Convergent validity was instead evaluated deriving the Pearson’s correlation coefficient or Spearman’s correlation coefficient, depending on data distribution. Comparisons between dominant and non-dominant trials were made with paired t tests or the Wilcoxon test, depending on the data distribution. Results from the test-retest reliability show an excellent, good or at least moderate ICC for all the kinematic parameters, indicating that the proposed protocol is consistent within the same session from one test administration to the next, where a single rater uses the same methods and the same testing conditions. This is a huge result and it indicates that adding new outcomes to the test leads to an improvement in the test-retest reliability in healthy subjects, in contrast with the low to moderate reliability of the standard NHPT (Mathiowetz, et al. 1985) (Grice, et al. 2003). Moreover, the statistical analysis results also an excellent, good or moderate intra-rater reliability for all kinematic outcomes, with the exception of the ROM of the trunk along X that shows a poor ICC. This result highlights that this instrumental version of the NHPT, despite the much more additional outcome variables with the respect to the standard NHPT, is however reliable and repeatable also between sessions, in response to the repositioning of the retroreflective markers by the same operator. The convergent validity analysis shows for most of the kinematic outcomes a very strong, moderate or fair correlation with the test total time. Only for a few parameters — time to peak, trunk ROMs and Indexes of Curvature — the correlation is poor, as a result of a limited variability of these kinematic parameters in healthy subjects, in agreement with literature (Johansson and Häger 2019). Lastly, from statistical analysis results, we can claim that this proposed protocol is able to discriminate significantly from dominant and non-dominant hand in most of its outcomes. This means that adding more variables than the standard one provides a better and more deeply characterization of subject’s manual dexterity. However, the test is not able to discriminate between right hand and left hand for mean velocity, trunk ROMs along X and Z and index of curvature outcomes, because of the limited variability in these parameters between the two hands in healthy subjects. Proving the good validity and reliability of this instrumental version of the NHPT allows to future applications in pathological subjects, where the increased variability of the outcome variables may enhance even more the importance to have much more valid and reliable kinematic parameters than the single outcome of the standard NHPT. Furthermore, this kinematic protocol is the first step towards a more complete and remarkable protocol, in which both kinematics and cortical activity are recorded simultaneously during the NHPT. Regarding this last protocol only a pilot study with few subjects was carried out in this dissertation, in order to briefly enlighten the pros and limitations of EEG integration with kinematics recording. The NHPT was completed in the exact same way of the first protocol, but the subject performed only one session (twice execution with right hand and twice with left hand). The only difference was the introduction of the EEG recording — using a brain cap with 61 channels — combined with optoelectronic system. In addition, in this pilot study, it was also explored cortical activity during action observation (AO) and motor imagery (MI) tasks, in order to study motor neuron activation during the observation or imagination of simple motor tasks of manual dexterity. Indeed, this work is part of a larger study that aims to demonstrate the effectiveness of action observation therapy (AOT) followed immediately by sleep, in order to improve and enhance manual dexterity in healthy subjects. Lately, there have been actually studies supporting and validating the efficacy of AO regarding both the improvement of a specific motor skill in healthy subjects and the rehabilitation of patients suffering from chronic ischemic stroke, PD patients, children with cerebral palsy, and non-neurological patients such as those undergoing orthopedic surgery of the hip or knee (Brunsdon, Bradford and Ferguson 2019) (Gatti, et al. 2013) (Zhu, et al. 2015). Concurrently, MI has been studied widely along with AO and it is also considered an effective tool (Tani, et al. 2017) (Marusic, et al. 2018). EEG pre-processing and data analysis were conducted with the collaboration of the research group of B3Lab – Biosignals, Bioimaging, Bioinformatics of the Politecnico di Milano. In particular, data’s noisy and irrelevant segments were manually eliminated and Independent Component Analysis (ICA) was applied to separate brain activity and noise sources. In order to determine the subject’s α frequency range, it was examined EEG recording during the eyes closed state when α rhythm is dominant, while the individual μ-rhythms needed the analysis of the open eyes’ baseline over the central channels that correspond to the motor area. Then, during the NHPT execution the relative power in the classical EEG frequency bands and the μ -band were extracted. Event-Related Spectral Perturbation (ERSP) maps were instead used to identify subject-specific frequency bands reactive to action observation and motor imagery (Tacchino, et al. 2016). Results from EEG recording during the manual motor task show that the μ rhythm, even if it is similar to the α rhythms, has two particularities: in the power maps during rest with eyes open the μ rhythm is more localized on the central channels with respect to the α power which spread posteriorly; moreover, a reduction of central μ power is clearly visible during movement. However, in this pilot study, we’ve encountered also some limitations in the integration of EEG system with the optoelectronic system. Some of them are purely technical problems regarding the experimental setup made tricky by the unwieldy of the EEG system that may interfere with subject’s movement or markers’ visibility, resulting sometimes in artifacts in the EEG signal. Others drawbacks concerns the fact that the NHPT is a simple and fast motor task and for this reason few outcome parameters can be derived from cortical activity. Nevertheless, some of these limitations may be easily overcome and this integration of EEG in kinematics recording can be a starting point and a solid basis for future investigation of the assessment of manual dexterity in both healthy and pathological subjects. On the other hand, from EEG recording during AO and MI tasks, we’ve found that during AO the desynchronization in the μ and β ranges is evident — even if the significance was not tested here —, whereas during MI this event is less observable. This distinction can be mostly observed in the ERSP images of C4, CP3 and CP4 electrodes, where the desynchronization is much more present during AO than MI. An explanation of that could be that during AO also visuo-motor system is involved, whereas in MI there isn’t a precise instant we can use to synchronize the EEG epoch since the subject can start imaging whenever she/he wants after the cue. This brief experimental trial has allowed us to point out the effectiveness of AO and MI of a simple motor task, since we can observe neuron activity also during observation or imagination. Furthermore, from this short study on AO and MI we’ve remarked that these innovative therapies can easily and efficiently be included in a period of training for healthy subjects to improve manual dexterity or rehabilitation for pathological patients to recover upper arm’s ability. In conclusion, in this dissertation we’ve carried out and demonstrate that the proposed instrumental version of the NHPT with kinematic recording has an acceptable test-rest reliability, intra-rater reliability, convergent validity and discriminant validity. For this reason, this protocol can be included in further applications in order to thoroughly assess manual dexterity in healthy subjects. Moreover, the same protocol may be in the future applicated and validated also with pathological patients. Moreover, in this work we’ve also mentioned and proposed an ulterior version of the instrumented NHPT, where not only kinematics but also cortical activity has been recorded. Despite some minor limitations, this innovative protocol may be a solid basis for future applications and a deeper understanding of manual dexterity both in healthy and pathological subjects. Lastly, we’ve also investigated cortical activity during AO and MI, reaching results in accordance with literature. All these studies are a starting point and a solid basis for a larger study that aims to demonstrate the effectiveness of AOT and MI followed immediately by sleep, in order to improve and enhance manual dexterity in healthy subjects.| File | Dimensione | Formato | |
|---|---|---|---|
|
Validation of the instrumental version of Nine Hole Peg Test in healthy subjects A starting point for EEG integration.pdf
non accessibile
Dimensione
4.03 MB
Formato
Adobe PDF
|
4.03 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/171084