Neurological disorders, including Spinal Cord Injury (SCI) and stroke, affect over one bil- lion people worldwide, severely impacting their sensorimotor functions and quality of life. The growing incidence of neurological disorders, coupled with the aging population, is causing an increase in the number of subjects who need rehabilitation therapies to restore physiological and motor functions that strongly impact the independence and community participation of affected individuals. Therefore, the optimization of such therapies is fundamental to maximize their efficacy. In this work, we will mainly focus on motor rehabilitation, specifically the one of the lower limbs, intended for gait training. Motor rehabilitation takes advantage of neuroplasticity through intense, task-oriented and repetitive exercises. This approach is widely recognized as the key mechanism driving functional improvements, especially during the initial phases following injury. Besides the Conventional Therapy approach, the integration of technologies within rehabilitation has gained increasing interest over time. On one hand, Rehabilitation Robotics has been introduced into clinical practice with the use of mechatronic devices for sup- porting or actuating limbs during therapy. These systems enable tailored, repetitive and challenging exercises, but are strongly limited by a heavy and bulky structure which de- creases their usability and acceptability from end-users. On the other hand, Electrical Stimulation (ES) has received particular attention over the past years, offering promising neuromodulation benefits for sensorimotor recovery in neurological disorders. It involves the delivery of low-energy electrical pulses to nerves located either peripherally or centrally. In the former case, it is termed Neuro-Muscular Electrical Stimulation (NMES), further categorized into Functional Electrical Stimulation (FES) and Sensory Afferent Electrical Stimulation (SAES). FES triggers muscle contractions to produce functional movements, thus it is administered at an intensity surpassing the motor threshold. SAES, instead, mainly focuses on sensory fibers for eliciting changes in sensorimotor functions. The latter alternative, instead, refers to Spinal Cord Stimulation (SCS), primarily employed to modulate muscle activity and alleviate spasticity. These ES-based strategies are associated with several peripheral and central advantages, depending on their specific nature. Nevertheless, they also face some limitations such as the high variability of stimulation-induced responses, both between subjects and within the same one, and the early appearance of muscle fatigue, particularly relevant when ES is delivered above the motor threshold. Some of these limitations can be overcome by combining ES-based neuromodulation techniques with robotics. This thesis aims to explore these different neuromodulation approaches, focusing on their working principle and induced neurological effects to discuss their potential application in rehabilitation settings. First, it wonders whether by exploiting FES-induced movement, it could be possible to reduce motors’ power in hybrid devices and, consequently, realize lighter devices. In the direction of improving the device usability, the possibility of using textile electrodes in place of hydrogel ones was also investigated, given their potential integration into clothes. Moreover, it inspects the neuromodulatory effects of SCS and the potential benefits of its combination with residual voluntary signals. Finally, the topic of afferent stimulation is expanded by comparing the reflex waves generated by SCS and SAES, inferring that they could elicit similar neuromodulatory effects. The first research question is addressed by integrating FES within a robotic device, realizing what is usually referred to as hybrid system. As encouraged by literature studies, this combination would enhance the advantages of the two technologies and mitigate their disadvantages. The addition of FES would directly engage the subject’s muscles in movement execution and reduce the motors’ contribution; simultaneously, the robotic support would be beneficial in delaying the onset of stimulation-induced muscle fatigue, a primary constraint in prolonged stimulation training. To this aim, the TwinFES prototype was realized by combining two commercial electrical stimulators into a motorized lower-limb exoskeleton (Twin, Italian Institute of Technology). The primary study focus has been the development of a cooperative control strategy that, integrating the two technologies, enables FES assistance to adapt and compensate for reductions in motor torque. An extensive validation of this hybrid control system was carried out on healthy subjects and individuals affected by either complete/incomplete SCI or stroke, recruited at the Villa Beretta rehabilitation center. Results of both studies confirmed the possibility of reducing motors’ power enabled by the inclusion of FES and proved that engaging subjects’ muscles in the movement execution increases the device’s usability. Nevertheless, the complexity of the overall system emerged as a primary limitation of the device. Current research is trying to tackle these shortcomings, such as the challenging fit and the need for precise electrode placement by therapists, aiming to present such systems as everyday assistive devices and increase users’ acceptance. In this direction, the possibility of using textile electrodes to improve the wearability of FES-based systems was explored. A novel set of textile electrodes was developed by integrating commercial screen-printed electrodes into textile strips and validated with respect to traditional hydrogel ones, in terms of electrical properties, overall stimulation comfort and performance on 14 healthy subjects. Collected results reported no significant differences among the two electrode types, confirming the possibility of substituting hydrogel samples with textile ones that, given their possible integration into clothes, would ease hybrid devices’ donning and doffing processes. Furthermore, the textile alternative would facilitate the transition from clinical-based therapies to home-based solutions. Then, the focus is shifted to sensory stimulation strategies, namely SCS and SAES, which act on afferent fibers to modulate the excitability of the central nervous system. Rather than inducing a movement, such stimulations aim to facilitate the possibility of residual weak signals to produce movements. To this aim, a protocol combining SCS with cycling was tested on four complete SCI individuals while recording the EMG signals of their lower-limbs muscles. Despite the limited number of performed sessions, preliminary but encouraging observations were retrieved from this study. These included a reduction of spasticity and some benefits on bladder and bowel function, which persisted for days after the stimulation. Direct motor facilitation was not reported, probably due to the limited number of performed sessions, but rather a promising SCS-dependent modulation of the EMG signal was observed. Finally, the investigation of afferent stimulation advances by comparing SCS and SAES, considering that both are based on a reflex-mediated activity. The study was motivated by the attempt to find an alternative solution to SCS, given its limitations: the impossibility of applying it in the presence of spinal metal implants, the difficulty of an autonomous placement of electrodes and the high intensity of required stimulation currents. Tests involved twelve healthy participants and compared the reflex waves induced by the two types of sensory ES in terms of shape, amplitude and latency. The study confirmed that the two types of stimulation techniques trigger the same spinal circuits but at different locations over the reflex arc. These preliminary results paved the way for future studies analyzing whether similar therapeutic benefits are prompted by the two strategies, in view of promoting their interchangeable use depending on the specific case. This work proposes the use of electrical stimulation therapies, and their possible integration with robotics, as a valid option for the rehabilitation of neurological diseases. It highlights the importance of defining the therapy’s intent to determine the appropriate stimulation strategy to be applied, with FES mainly employed for the induction of functional movements and sensory stimulation (SCS and SAES) mainly intended for neuromodulation. By sending afferent signals to the brain, both alternatives were also demonstrated to promote plasticity processes and motor learning, particularly when combining the proprioceptive feedback from stimulation with volitional efforts. On one hand, this study proves the efficacy of cooperative FES-motor control strategies, particularly in reducing motor torque demands and prolonging benefits of FES, reducing muscle fatigue. Improvements in system usability were also observed, which could be further enhanced by replacing hydrogel electrodes with textile ones, given their demonstrated equivalent properties. On the other hand, it explores the therapeutic benefits of SCS and offers a preliminary investigation of SAES, emphasizing the need for further inspection of its therapeutic impact. Challenges and limitations of these techniques are also discussed, prompting further research to develop more advanced and flexible systems aimed at im- proving rehabilitation outcomes.
I disturbi neurologici, inclusi la lesione del midollo spinale (SCI) e l'ictus, colpiscono oltre un miliardo di persone in tutto il mondo, compromettendo gravemente le loro funzioni sensoriali e motorie e la qualità della vita. La crescente incidenza di disturbi neurologici, insieme all'invecchiamento della popolazione, sta causando un aumento del numero di soggetti che necessitano di terapie riabilitative per ripristinare le funzioni fisiologiche e motorie, influenzando fortemente l'indipendenza e la partecipazione comunitaria degli individui colpiti. Pertanto, l'ottimizzazione di tali terapie è fondamentale per massimizzarne l'efficacia. In questo lavoro ci concentreremo principalmente sulla riabilitazione motoria, in particolare quella degli arti inferiori, destinata all'allenamento del cammino. La riabilitazione motoria sfrutta la neuroplasticità attraverso esercizi intensi, funzionali e ripetitivi. Questo approccio è ampiamente riconosciuto come il meccanismo chiave che guida i miglioramenti funzionali, soprattutto nelle fasi iniziali successive all'infortunio. Oltre alla Terapia Convenzionale, l'integrazione delle tecnologie nella riabilitazione ha guadagnato crescente interesse nel tempo. Da un lato, la Robotica Riabilitativa è stata introdotta nella pratica clinica con l'uso di dispositivi meccatronici per supportare il movimento durante la terapia. Questi sistemi consentono esercizi personalizzati, ripetitivi e impegnativi, ma sono fortemente limitati da una struttura pesante e ingombrante che ne riduce l'usabilità e l'accettabilità da parte degli utenti. Dall'altro lato, la Stimolazione Elettrica (ES) ha ricevuto particolare attenzione negli ultimi anni, offrendo promettenti benefici di neuromodulazione per il recupero sensomotorio nei disturbi neurologici. Essa prevede la somministrazione di impulsi elettrici a bassa energia ai nervi, sia periferici che centrali. Nel primo caso, è definita Stimolazione Elettrica Neuro-Muscolare (NMES), ulteriormente categorizzata in Stimolazione Elettrica Funzionale (FES) e Stimolazione Elettrica Afferente Sensoria (SAES). La FES induce contrazioni muscolari per produrre movimenti funzionali, quindi viene somministrata ad un'intensità superiore alla soglia motoria. La SAES, invece, si concentra principalmente sulle fibre sensoriali per indurre cambiamenti nelle funzioni sensomotorie. L'altra alternativa è rappresentata dalla Stimolazione del Midollo Spinale (SCS), impiegata principalmente per modulare l'attività muscolare e alleviare la spasticità. Queste strategie basate sulla stimolazione sono associate a diversi vantaggi periferici e centrali, a seconda della loro specifica natura. Tuttavia, presentano anche alcune limitazioni come l'elevata variabilità delle risposte indotte dalla stimolazione, sia tra i soggetti che all'interno dello stesso soggetto, e la precoce comparsa dell'affaticamento muscolare, particolarmente rilevante quando la stimolazione viene somministrata al di sopra della soglia motoria. Alcune di queste limitazioni possono essere superate combinando le tecniche di neuromodulazione con la robotica. Questa tesi mira a esplorare questi diversi approcci di neuromodulazione, concentrandosi sul loro principio di funzionamento e sugli effetti neurologici indotti per discutere la loro potenziale applicazione nei contesti riabilitativi. In primo luogo, si indaga se il movimento indotto dalla FES permetta di ridurre la potenza dei motori nei dispositivi ibridi e, di conseguenza, realizzare dispositivi più leggeri. Per migliorare l'usabilità del dispositivo, è stata poi investigata anche la possibilità di utilizzare elettrodi tessili al posto di quelli in gel, data la loro potenziale integrazione nei vestiti. Inoltre, si analizzano gli effetti neuromodulatori della SCS e i potenziali benefici della sua combinazione con i segnali volontari residui. Infine, il tema della stimolazione afferente viene ampliato confrontando le onde riflesse generate da SCS e SAES, ipotizzando che possano indurre effetti neuromodulatori simili. La prima domanda di ricerca viene affrontata integrando la FES all'interno di un dispositivo robotico, realizzando quello che viene solitamente definito sistema ibrido. Come incoraggiato dagli studi in letteratura, questa combinazione accentuerebbe i vantaggi delle due tecnologie e ne mitigherebbe gli svantaggi. L'aggiunta della FES coinvolgerebbe direttamente i muscoli del soggetto nell'esecuzione del movimento e ridurrebbe il contributo dei motori; contemporaneamente, il supporto robotico sarebbe benefico nel ritardare l'insorgenza dell'affaticamento muscolare indotto dalla stimolazione, un vincolo primario nell'allenamento con stimolazione prolungata. A tal fine, è stato realizzato il prototipo TwinFES combinando due stimolatori elettrici commerciali in un esoscheletro motorizzato per arti inferiori (Twin, Istituto Italiano di Tecnologia). Il focus principale dello studio è stato lo sviluppo di una strategia di controllo cooperativa che, integrando le due tecnologie, consente all'assistenza della FES di adattarsi e compensare le riduzioni della coppia ai motori. È stata effettuata un'estesa validazione di questo sistema di controllo ibrido su soggetti sani e individui affetti da SCI completo/incompleto o ictus, reclutati presso il centro di riabilitazione Villa Beretta. I risultati di entrambi gli studi hanno confermato la possibilità di ridurre la potenza dei motori grazie all'inclusione della FES e hanno dimostrato che il coinvolgimento dei muscoli dei soggetti nell'esecuzione del movimento aumenta l'usabilità del dispositivo. Tuttavia, la complessità dell'intero sistema è emersa come una limitazione primaria del dispositivo. La ricerca attuale sta cercando di affrontare queste carenze, come la difficile vestibilità e la necessità di un posizionamento preciso degli elettrodi da parte dei terapisti, con l'obiettivo di presentare tali sistemi come dispositivi assistivi quotidiani e aumentarne l'accettazione da parte degli utenti. In questa direzione, è stata esplorata la possibilità di utilizzare elettrodi tessili per migliorare la vestibilità dei sistemi basati sulla FES. Un nuovo set di elettrodi tessili è stato sviluppato integrando elettrodi commerciali stampati con inchiostro conduttivo in strisce tessili e validato rispetto a quelli tradizionali in gel, in termini di proprietà elettriche, comfort complessivo della stimolazione e prestazioni su 14 soggetti sani. I risultati raccolti non hanno riportato differenze significative tra i due tipi di elettrodi, confermando la possibilità di sostituire gli elettrodi a gel con quelli tessili che, data la loro possibile integrazione nei vestiti, faciliterebbero i processi di indossamento e svestizione dei dispositivi ibridi. Inoltre, l'alternativa tessile faciliterebbe lo spostamento delle terapie dall’ambiente clinico a quello domestico. Successivamente, l'attenzione si è spostata sulle strategie di stimolazione sensoriale, in particolare SCS e SAES, che agiscono sulle fibre afferenti per modulare l'eccitabilità del sistema nervoso centrale. Piuttosto che indurre un movimento, tali stimolazioni mirano a facilitare la possibilità di segnali residui deboli di produrre movimenti. A tal fine, è stato testato un protocollo che combina SCS con la pedalata su quattro individui con SCI completo registrando i segnali EMG dei loro muscoli degli arti inferiori. Nonostante il numero limitato di sessioni eseguite, questo studio ha fornito osservazioni preliminari ma incoraggianti. In particolare, una riduzione della spasticità e alcuni benefici sulle funzioni della vescica e dell'intestino, che persistevano per giorni dopo la stimolazione. Non è stata riportata una facilitazione motoria diretta, probabilmente a causa del numero limitato di sessioni eseguite, ma piuttosto è stata osservata una promettente modulazione del segnale EMG dipendente dalla SCS. Infine, l'indagine sulla stimolazione afferente ha confrontato SCS e SAES, considerando che entrambe si basano su un'attività riflessa. Lo studio è stato motivato dal tentativo di trovare una soluzione alternativa alla SCS, date le sue limitazioni: l'impossibilità di applicarla in presenza di impianti metallici spinali, la difficoltà di un posizionamento autonomo degli elettrodi e l'elevata intensità delle correnti di stimolazione richieste. I test hanno coinvolto dodici partecipanti sani e hanno confrontato le onde riflesso indotte dai due tipi di stimolazione sensoriale in termini di forma, ampiezza e latenza. Lo studio ha confermato che le due tecniche di stimolazione attivano gli stessi circuiti spinali ma in punti diversi lungo l'arco riflesso. Questi risultati preliminari hanno aperto la strada a futuri studi per analizzare se i benefici terapeutici indotti dalle due strategie siano simili, con l'obiettivo di promuoverne l'uso intercambiabile a seconda del caso specifico. Questo lavoro propone l'uso delle terapie di stimolazione elettrica, e la loro possibile integrazione con la robotica, come una valida opzione per la riabilitazione delle malattie neurologiche. Sottolinea l'importanza di definire l'intento della terapia per determinare la strategia di stimolazione appropriata da applicare, con la FES utilizzata principalmente per l'induzione di movimenti funzionali e la stimolazione sensoriale (SCS e SAES) destinata principalmente alla neuromodulazione. Inviando segnali afferenti al cervello, entrambe le alternative hanno dimostrato di promuovere processi di plasticità e apprendimento motorio, in particolare quando si combina il feedback propriocettivo della stimolazione con gli sforzi volontari. Da un lato, questo studio dimostra l'efficacia delle strategie di controllo cooperativo FES-motore, in particolare nella riduzione delle richieste di coppia motoria e nel prolungare i benefici della FES, riducendo l'affaticamento muscolare. Sono stati anche osservati miglioramenti nell'usabilità del sistema, che potrebbero essere ulteriormente aumentati sostituendo gli elettrodi in gel con quelli tessili, date le loro dimostrate proprietà equivalenti. Dall'altro lato, esplora i benefici terapeutici della SCS e offre un'indagine preliminare della SAES, sottolineando la necessità di ulteriori ricerche per approfondire il suo impatto terapeutico. Sono inoltre discusse le sfide e i limiti di queste tecniche, promuovendo ulteriori ricerche per sviluppare sistemi più avanzati e flessibili che mirano a migliorare i risultati della riabilitazione.
Neuromodulation approaches for the rehabilitation of neurological diseases : Noninvasive Electrical Stimulation and Robotics
Dell'EVA, FRANCESCA
2023/2024
Abstract
Neurological disorders, including Spinal Cord Injury (SCI) and stroke, affect over one bil- lion people worldwide, severely impacting their sensorimotor functions and quality of life. The growing incidence of neurological disorders, coupled with the aging population, is causing an increase in the number of subjects who need rehabilitation therapies to restore physiological and motor functions that strongly impact the independence and community participation of affected individuals. Therefore, the optimization of such therapies is fundamental to maximize their efficacy. In this work, we will mainly focus on motor rehabilitation, specifically the one of the lower limbs, intended for gait training. Motor rehabilitation takes advantage of neuroplasticity through intense, task-oriented and repetitive exercises. This approach is widely recognized as the key mechanism driving functional improvements, especially during the initial phases following injury. Besides the Conventional Therapy approach, the integration of technologies within rehabilitation has gained increasing interest over time. On one hand, Rehabilitation Robotics has been introduced into clinical practice with the use of mechatronic devices for sup- porting or actuating limbs during therapy. These systems enable tailored, repetitive and challenging exercises, but are strongly limited by a heavy and bulky structure which de- creases their usability and acceptability from end-users. On the other hand, Electrical Stimulation (ES) has received particular attention over the past years, offering promising neuromodulation benefits for sensorimotor recovery in neurological disorders. It involves the delivery of low-energy electrical pulses to nerves located either peripherally or centrally. In the former case, it is termed Neuro-Muscular Electrical Stimulation (NMES), further categorized into Functional Electrical Stimulation (FES) and Sensory Afferent Electrical Stimulation (SAES). FES triggers muscle contractions to produce functional movements, thus it is administered at an intensity surpassing the motor threshold. SAES, instead, mainly focuses on sensory fibers for eliciting changes in sensorimotor functions. The latter alternative, instead, refers to Spinal Cord Stimulation (SCS), primarily employed to modulate muscle activity and alleviate spasticity. These ES-based strategies are associated with several peripheral and central advantages, depending on their specific nature. Nevertheless, they also face some limitations such as the high variability of stimulation-induced responses, both between subjects and within the same one, and the early appearance of muscle fatigue, particularly relevant when ES is delivered above the motor threshold. Some of these limitations can be overcome by combining ES-based neuromodulation techniques with robotics. This thesis aims to explore these different neuromodulation approaches, focusing on their working principle and induced neurological effects to discuss their potential application in rehabilitation settings. First, it wonders whether by exploiting FES-induced movement, it could be possible to reduce motors’ power in hybrid devices and, consequently, realize lighter devices. In the direction of improving the device usability, the possibility of using textile electrodes in place of hydrogel ones was also investigated, given their potential integration into clothes. Moreover, it inspects the neuromodulatory effects of SCS and the potential benefits of its combination with residual voluntary signals. Finally, the topic of afferent stimulation is expanded by comparing the reflex waves generated by SCS and SAES, inferring that they could elicit similar neuromodulatory effects. The first research question is addressed by integrating FES within a robotic device, realizing what is usually referred to as hybrid system. As encouraged by literature studies, this combination would enhance the advantages of the two technologies and mitigate their disadvantages. The addition of FES would directly engage the subject’s muscles in movement execution and reduce the motors’ contribution; simultaneously, the robotic support would be beneficial in delaying the onset of stimulation-induced muscle fatigue, a primary constraint in prolonged stimulation training. To this aim, the TwinFES prototype was realized by combining two commercial electrical stimulators into a motorized lower-limb exoskeleton (Twin, Italian Institute of Technology). The primary study focus has been the development of a cooperative control strategy that, integrating the two technologies, enables FES assistance to adapt and compensate for reductions in motor torque. An extensive validation of this hybrid control system was carried out on healthy subjects and individuals affected by either complete/incomplete SCI or stroke, recruited at the Villa Beretta rehabilitation center. Results of both studies confirmed the possibility of reducing motors’ power enabled by the inclusion of FES and proved that engaging subjects’ muscles in the movement execution increases the device’s usability. Nevertheless, the complexity of the overall system emerged as a primary limitation of the device. Current research is trying to tackle these shortcomings, such as the challenging fit and the need for precise electrode placement by therapists, aiming to present such systems as everyday assistive devices and increase users’ acceptance. In this direction, the possibility of using textile electrodes to improve the wearability of FES-based systems was explored. A novel set of textile electrodes was developed by integrating commercial screen-printed electrodes into textile strips and validated with respect to traditional hydrogel ones, in terms of electrical properties, overall stimulation comfort and performance on 14 healthy subjects. Collected results reported no significant differences among the two electrode types, confirming the possibility of substituting hydrogel samples with textile ones that, given their possible integration into clothes, would ease hybrid devices’ donning and doffing processes. Furthermore, the textile alternative would facilitate the transition from clinical-based therapies to home-based solutions. Then, the focus is shifted to sensory stimulation strategies, namely SCS and SAES, which act on afferent fibers to modulate the excitability of the central nervous system. Rather than inducing a movement, such stimulations aim to facilitate the possibility of residual weak signals to produce movements. To this aim, a protocol combining SCS with cycling was tested on four complete SCI individuals while recording the EMG signals of their lower-limbs muscles. Despite the limited number of performed sessions, preliminary but encouraging observations were retrieved from this study. These included a reduction of spasticity and some benefits on bladder and bowel function, which persisted for days after the stimulation. Direct motor facilitation was not reported, probably due to the limited number of performed sessions, but rather a promising SCS-dependent modulation of the EMG signal was observed. Finally, the investigation of afferent stimulation advances by comparing SCS and SAES, considering that both are based on a reflex-mediated activity. The study was motivated by the attempt to find an alternative solution to SCS, given its limitations: the impossibility of applying it in the presence of spinal metal implants, the difficulty of an autonomous placement of electrodes and the high intensity of required stimulation currents. Tests involved twelve healthy participants and compared the reflex waves induced by the two types of sensory ES in terms of shape, amplitude and latency. The study confirmed that the two types of stimulation techniques trigger the same spinal circuits but at different locations over the reflex arc. These preliminary results paved the way for future studies analyzing whether similar therapeutic benefits are prompted by the two strategies, in view of promoting their interchangeable use depending on the specific case. This work proposes the use of electrical stimulation therapies, and their possible integration with robotics, as a valid option for the rehabilitation of neurological diseases. It highlights the importance of defining the therapy’s intent to determine the appropriate stimulation strategy to be applied, with FES mainly employed for the induction of functional movements and sensory stimulation (SCS and SAES) mainly intended for neuromodulation. By sending afferent signals to the brain, both alternatives were also demonstrated to promote plasticity processes and motor learning, particularly when combining the proprioceptive feedback from stimulation with volitional efforts. On one hand, this study proves the efficacy of cooperative FES-motor control strategies, particularly in reducing motor torque demands and prolonging benefits of FES, reducing muscle fatigue. Improvements in system usability were also observed, which could be further enhanced by replacing hydrogel electrodes with textile ones, given their demonstrated equivalent properties. On the other hand, it explores the therapeutic benefits of SCS and offers a preliminary investigation of SAES, emphasizing the need for further inspection of its therapeutic impact. Challenges and limitations of these techniques are also discussed, prompting further research to develop more advanced and flexible systems aimed at im- proving rehabilitation outcomes.File | Dimensione | Formato | |
---|---|---|---|
DellEvaFrancesca_PhDThesis.pdf
accessibile in internet per tutti
Descrizione: PhD thesis of Francesca Dell'Eva
Dimensione
20.02 MB
Formato
Adobe PDF
|
20.02 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/10589/222434