"A Wearable Ring-shaped Inertial System to Identify Action Planning Impairments During Reach-to-grasp Sequences: A Pilot Study"

doi:10.21203/rs.3.rs-118987/v1

Download PDF

Research

"A Wearable Ring-shaped Inertial System to Identify Action Planning Impairments During Reach-to-grasp Sequences: A Pilot Study"

https://doi.org/10.21203/rs.3.rs-118987/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 27 Jul, 2021

Read the published version in Journal of NeuroEngineering and Rehabilitation →

You are reading this latest preprint version

Background. The progressive ageing of the population is leading to an increasing of people affected by cognitive decline, included disorders in executive functions (EFs), such as action planning. Current procedures to evaluate cognitive decline are based on neuropsychological tests, but novel methods and approaches start to be investigated. Reach-to-grasp (RG) protocols have shown that intentions can influence the EFs of action planning.

Methods. In this work, we proposed a novel ring-shaped wearable inertial device, SensRing, to measure kinematic parameters during RG and after-grasp tasks with different end-goals. The aim is to evaluate whether SensRing can characterize the motor performances of people affected by Mild Neurocognitive Disorder (MND) with impairment in EFs, named dysexecutive MND (d-MND), compared to older healthy subjects (HC). Ten HC and eight d-MND were asked to reach and grasp a can with three different intentions: to drink (DRINK), to place it on a target (PLACE), or to pass it to a partner (PASS). Twenty-one kinematic parameters have been extracted from SensRing inertial data.

Results. Several parameters resulted able to differentiate between HC and d-MND, showing that patients generally have longer reaction times, slower peak velocities and longer deceleration phases. Furthermore, d-MND showed fewer differences among conditions, suggesting that impairments in EFs influence their capabilities into modulating the action planning based on the end-goal.

Conclusions. Therefore, the system can be proposed for objectifying the clinical assessment of people affected by d-MND by administering an easy quick motor test. Furthermore, the portability, wearability, accuracy, and ease-of use of the system make the SensRing sytem suitable for remote applications at home, including analysis of protocols for neuromotor rehabilitation in patients affected by MND.

Physical Medicine & Rehab

Action planning

Dysexecutive Syndrome

Executive Functions

Mild cognitive impairment

Motor Programming

Reach-to-grasp

The progressive ageing of the global population is leading to an increasing number of people affected by cognitive decline and dementia [1]. Particularly, it is expected that the number of people suffering from AD (accounting for 60–65% of the dementia cases) will reach 74.7 million by 2030 and 100 million by 2050 [2]. Even though dementia is mainly associated to the prototypic memory loss, different cognitive domains can be affected by different pathologies, leading to distinct cognitive symptoms. Among them, the executive functions (EFs) represent a complex construct that involves cognitive, behavioural, and emotional aspects. Deficits in EFs have been, so far, defined as “frontal syndrome” referring to the neuro-anatomical regions involved. Lately, it has been preferred the nomenclature “dysexecutive syndrome” that particularly refers to the functional decline of the superior executive functions [3]. Based on diagnostic criteria, dysexecutive syndrome includes cognitive (e.g., deficits in response inhibition, rules deduction, set-shifting, information generation, action planning, response initiation, coordination of dual-tasks) and/or behavioural (e.g., hypoactivity, apathy, distractibility, preservative behaviour, social behaviour) alterations [4].

Currently, EFs are clinically evaluated mainly administering standardized neuropsychological tests. In particular, different tests are traditionally applied to investigate the EFs and their decline considering several aspects [5]. For instance, the Frontal Assessment Battery (FAB) and the Behavioral Assessment of Dysexecutive Syndrome are the most used tests to evaluate the EFs as a whole. Tests assessing specific aspects of EFs do also exist, such as the Trail Making Test (TMT), mainly assessing divided attention and the executive center of the working memory; the Stroop Interference Test, focused on the response inhibition; the Digit Span, to test the slave systems for the working memory (verbal side); the Tower of London addressing the planning capability; and the Hayling Sentence Completion Task investigating response initiation. A description of these and other commonly used tests has been reported by Dirmberger and Jahanshahi [5]. Although neuropsychological testing is today the gold standard to assess dysexecutive symptoms, a recent literature review highlighted that they present several limitations. The validity and reliability of the test results are sometimes limited because of normative data based on small datasets, some of the cognitive domains are scarcely represented, while others are assessed in different tasks, many tests are available in a restricted number of languages, and sometimes cultural habits can affect the execution of the required tasks [6].

In this context, new protocols and novel tools to assess neuropsychological functions should be investigated. These solutions ought to support the clinicians in objectively detecting cognitive declines, related to specific cognitive domains, both treatment and/or prevention improvement. In this work, we focus on the decline of EFs in motor programming that results in action planning impairments.

In the past years, some experimental studies, aiming to highlight how the intentions can influence the action planning, have proposed reach-to-grasp (RG) protocols. RG sequences analyses have revealed that healthy subjects differently reach and grasp an object depending on the action final goal [7]. When we perform a movement, we are driven by prior intentions, that change according to the action end-goal. Kinematics conveys information about these intentions [8], so that, even if the object to-be-grasped is the same, different motor parameters can be appreciated [9]. Therefore, when someone reaches and grasp a bottle to pour its content into a container or, conversely, to pass it to someone else, modulation of the kinematic action occurs.

Some works have evidenced that the ageing process can affect motor performances during RG protocols. Particularly, the older adults movements result slower, longer, and show a prolonged approach phase to the object [10]. This movement protraction can be explained as an increment of time that older persons need to develop compensative strategies while to cope with the dexterity loss, and operational capabilities deterioration [11]. However, protocols which involve highly goal-directed tasks seem to mitigate the differences between older adults and controls [12]. Furthermore, previous studies that implemented experimental protocols based on RG and after-grasp (AG) sequences have revealed useful information in several pathologies, such as Parkinson’s Disease [9], [13], autism spectrum disorder [14], [15], and stroke [16].

Reach-to-grasp tests have shown considerable advantages. For instance, they are easy to be performed and can overcome languages and cultural bias. Nevertheless, up to now, the traditional methodologies employed to analyze motor performance during such tasks are mainly based on motion optical capture systems, which are expensive, require lengthy procedures and dedicated staff for set-up and analysis, and are appliable in dedicated wide settings only.

Recently, advances in Micro Electro-Mechanical Systems (MEMS), and in artificial intelligence (AI) techniques, have allowed employing wearable technology, together with processing and learning algorithms, to evaluate experimental protocols, both in lab and in clinical settings. This represents a promising solution for objective and reliable monitoring, assessment, and support [17]. Hence, wearable inertial devices have been used, so far, to acquire and process high-frequency rate data to analyze motion performances within several applications, including daily activity living gestures [18], early Alzheimer’s detection [19], Parkinson’s disease assessment [20], mild cognitive impairment evaluation [21], and autism spectrum disorder assistance [22].

In this context, we propose SensRing: a non-invasive, low-cost, lightweight, easy-to-use, ergonomic device able to capture the 3D movements of a finger in the space. In previous work, we evaluated the accuracy of this device in measuring kinematic parameters in healthy people during standard exercises [23]. Here, the device is proposed within a pilot study, for the usage in a clinical application with people suffering from mild cognitive decline and healthy controls. The SensRing allows the measurement of the kinematic parameters related to the motor performance without interfering on motion capabilities.

This study proposes the use of SensRing as an alternative approach to traditional methods aiming to objectively analyze RG and AG sequences. Since Mild Neurocognitive Disorder (MND) subjects, with a decline in EFs, often show impairment in motor programming and action planning [24], the cognitive decline could be objectively confirmed through kinematic parameters’ variations. The hypothesis is that the cognitive decline could be identified objectively examining the kinematic parameters. The idea is to investigate: i) whether the kinematic performance of people diagnosed as MND with EF decline is different compared to older healthy controls, during a simple motor protocol, and ii) whether there are differences in action kinematic modulation depending on the action end-goal between MND and healthy subjects. Finally, we evaluated whether the kinematic parameters could be correlated to the traditional cognitive assessment based on clinical scores.

The possibility to objectify the assessment of the tasks, complementing the neuropsychological testing through finer, quantitative measures related to motion performance could be a groundbreaking achievement in the field of early diagnosis of dementia, able to significantly improve the characterization of MND subjects.

To the best of our knowledge, this is the first study that employes a non-invasive wearable ring-shaped device to measure the kinematics of the movement during RG sequences, aiming to evaluate whether differences in motor performances can characterize people affected by MND with a specific impairment of EFs.

A. Participants

Ten healthy controls (HC) (6 females, 4 males, mean age ± standard deviation (SD) 63.7 ± 9.9 years old) and 8 subjects diagnosed as Mild Neurocognitive Disorder (MND) (7 females, 1 male, mean age ± SD 75.7 ± 5.5 years old) were recruited at the Nice Research Memory Center (CMRR) & Cognition Behaviour Technology laboratory (CoBTeK). All the subjects were recruited in the context of Marco-Sense multi-centric research protocol. The study was performed in accordance with the Declaration of Helsinki and approved by the ethical committee CPP Ile de France (N° IDRCB: 2019-A00342-55). All participants received detailed written explanations on the study and signed written informed consent. Participants were not included in the study if they had a score at the Mini-Mental State Examination (MMSE) < 22 [25] and a Frontal Assessment Battery (FAB) score < 11 [26]. All the participants were right-handed. Eventually, based on the Bank National Alzheimer Diagnosis, all the MND patients have been classified according to the predominant deficit. Particularly, MND subjects which present dysexecutive MND impairments (e.g. action planning and motor programming deficits) have been named as d-MND.

B. Instruments

A novel ring-shaped device, called SensRing (Fig. 1), has been developed at the BioRobotics Institute of Scuola Superiore Sant’Anna (Pisa, Italy) to fully track the orientation and movement of the finger where it is worn [23]. The device, based on an ARM®Cortex™-M3 32-bit STM32-F103 microcontroller (STMicroelectronics, Italy), can acquire and store data with 50 Hz sampling frequency. SensRing mount 9–axes inertial measurement unit (IMU) LSM9DS1 (STMicroelectronics, Italy), including a 3D digital linear acceleration sensor (selectable full scale: ±2/±4/±8/±16 g), 3D digital angular rate sensor (selectable full scale: ±245/±500/±2000 dps), and 3D digital magnetic sensor (selectable full scale: ±4/±8/±12/±16 gauss). SensRing selected 2 g, 2000 dps, and 4 gauss as full scales. An integrated Bluetooth module (Rigado BMD-350, Nordic Semiconductor, Norway) allows wireless communication for data transmission towards a generic control station. A dedicated interface, developed in Visual Studio 2019 (Microsoft Corporation, USA) and based on C# language, ensures managing the connection and the acquisition of sensors data. A small, rechargeable PoLi battery, externally fixed to the wrist with an elastic band, supplies the SensRing. Integrated solutions for the battery are currently under development.

C. Experimental Protocol

Before starting the trial, the experimenters required the subjects to wear SensRing on the proximal phalanx of the dominant index finger. Participants sat in front of a table, laying their dominant hand on the starting position (3 cm from the edge of the table midsagittal position, and 15 cm away from the midsection). Experimenters instructed the participants to perform shorts reach-to-grasp (RG) and after-grasp (AG) sequences with three different end-goals, adapted from a previous study [7]. For each task, a can (ø = 5 cm, h = 8.5 cm) has been positioned in front of the participant, at 21 cm from the hand starting position along the midsagittal plane. The initial position was acquired as a baseline for 5 seconds. Afterwards, a tone was indicated the beginning of the task. For each condition, 10 repetitions have been performed. During the drinking condition (DRINK), subjects had to reach the can, grasp it, and lift it up simulating a drinking action. On the other hand, during the placing condition (PLACE), subjects had to reach the can, grasp it and place it inside a cup (ø = 7 cm), located 28 cm at the right side with respect to the initial position of the can. Finally, the passing condition (PASS) required the subjects to reach the can, grasp it and pass it to a partner. The partner sat to the right side of the table with the hand resting (on the same position of the cup in the previous condition) ready to take the can. Both for PLACE and PASS, the can was repositioned on the initial position after each repetition. The order of conditions was randomized across participants.

D. Signal Processing and Feature Extraction

Inertial data acquired with SensRing have been stored and offline processed by using MATLAB R2018a (The MathWorks, Inc., Natick, MA, USA). Accelerations and angular velocities, acquired from the accelerometer and gyroscope integrated into the IMU, were pre-processed with a fourth-order low-pass digital Butterworth filter, using a 5 Hz cut-off frequency to delete high-frequency noise.

Two characteristic phases have been identified in each repetition: the RG phase from the beginning of the action to the grasping of the object, and the AG phase from the object grasping to the end of the task. Accordingly, custom algorithms have been implemented to segment the signal into these phases, across each exercise. The dominant axis of the angular velocity has been used as the reference signal for the segmentation, and three characteristic times have been calculated for each repetition: the starting time, the grasping time, and the end time (Fig. 2). Then, a set of kinematic parameters was computed from accelerations and angular velocities as detailed in Table 1, to investigate different aspects of the motor performances including, for instance, energy, duration, velocity [27]–[29].

All the parameters have been measured for each repetition both in the RG and the AG phases, except for the reaction time, the amplitude and time of maximum hand opening that have been calculated during the RG phase only. Totally, 21 parameters composed the dataset of each exercise (i.e., 12 features for the RG and 9 for the AG phase).

E. Data Analysis

Qualitative variables, such as gender and education level, have been compared using Chi2 test, whereas the remaining clinical data measured, as quantitative variables, through the Mann-Whitney U-test. The Kolmogorov-Smirnov test has been preliminarily applied to verify the data distribution of each extracted parameter (see Table 1). Since all parameters resulted as not normally distributed, non-parametric statistical tests have been adopted for data analysis.

Specifically, two macro-analyses were carried out:

1) Inter-group analysis: to evaluate if kinematic parameters may be able to differentiate HC and d-MND in terms of action planning and performance. The non-parametric Mann-Whitney U-test was applied to investigate significant differences (p < 0.05) between the two groups.

2) Intra-group analysis: to investigate, within each group, if motor patterns are modulated based on the action planning and the execution of three conditions with different end-goal. Non-parametric Wilcoxon test was used to find significant differences (significant level set at p-value < 0.05) between pairs of conditions (i.e., DRINK vs PLACE, DRINK vs PASS, PLACE vs PASS).

Additionally, a correlation analysis of the motor performance to the MMSE score has been executed for each parameter, calculating the Spearman’s correlation coefficients. This analysis investigates whether the motor parameters correlate to the clinical score of a standard neuropsychological test, typically used as a screening tool for cognitive assessment. d-MND and HC are considered as a unique group for this analysis.

Subjects demographic and clinical characteristics have been reported in Table 2. A significant statistical difference on age has been attested between the two groups. Conversely, no differences emerged regarding the educational level and the gender of participants. Moreover, as expected, the two groups differ in terms of the global level of cognitive functioning, according to the MMSE score (p<0.0001).

A. Intergroup analysis

Healthy subjects differ from d-MND patients in several kinematic parameters. Summarizing, 7 parameters differentiate the two groups in the RG phase, while 8 are significant in the AG phase. Results from RG phase are reported in Table 3 for all the conditions (DRINK, PLACE, or PASS), whereas complete results for AG phase are shown in Table 4.

Trends characterizing the motor performance of the two groups are clearly recognizable. For RG, HC always reached a higher peak velocity compared to d-MND, performing faster movements in all the conditions (significant in PLACE and PASS), confirmed by an execution time longer in d-MND compared to HC. Moreover, the reaction times in d-MND subjects are longer than HC (significant in PLACE). The d-MND subjects always showed a longer deceleration phase (significant in PASS) that is confirmed by the relative value with respect to the entire execution time of the action (T_decel_perc significant in DRINK and PLACE). Also, the lower Rmse_Jerk in patients compared to HC (except for PLACE where the values are almost identical), might be caused by lower accelerations and slower movements in d-MND. Additionally, d-MNDs have higher kurtosis than HC, showing a larger distribution data; also, the higher skewness in d-MND (except for PASS), confirms the wider variability in performing repetitive tasks in patients respect to HC. Finally, d-MND have higher IAV than HC, which can be referred to a higher request of effort to perform the same action.

Even if the RG phase is the most significant to investigate impairments in action planning, the same trends are found also in the AG phase for IAV, RMSE_Jerk, skewness, kurtosis, Vpeak, T_exec, T_decel, T_decel_perc (see Table 4), enhancing the reliability of these parameters in identifying motor impairments in patients affected by the dysexecutive syndrome.

Interestingly, the HC group show higher amplitude in hand opening when they must grasp an object; additionally, the maximum opening of their hand happens earlier than d-MND.

B. Intragroup analysis

Comparing the three tasks with different end-goals within each group, several significant differences come out. It is important investigating this comparison in the RG phase, which is the same for all the tasks.

As expected, HC showed more differences among conditions compared to d-MND in RG. Significant differences revealed in d-MND are confirmed also in HC; moreover, healthy subjects showed additional 9 parameters able to differentiate among tasks, showing that they modulated more consistently their kinematics according to the different action end-goal (Fig. 3). Particularly, d-MND never show significant differences in distinguishing between PLACE and PASS.

IAV, kurtosis, and T_exec parameters have an increasing trend in values from DRINK, through PLACE, to PASS, both for d-MND and for HC. The passing condition, indeed, seems to be the most demanding task, showing the highest IAV value, accompanied by the highest kurtosis, and the longest execution time. Conversely, drinking appears the simplest task, requiring less efforts and shorter times. Furthermore, DRINK has a deceleration phase (T_decel, T_decel_perc) lower than the other two conditions for both groups.

Interestingly, Vpeak is highest in PLACE condition, both for d-MND and HC. Whereas, a significant difference has found for skewness in HC comparing PASS to PLACE. Notably, the different end-goal seems not to influence the kinematic of the hand opening, neither the time when the hand reaches the maximum excursion approaching the can.

C. Motor Parameters Correlation to Clinical Score

Analyzing the correlation between the measured parameters and clinical score assigned to each participant (both d-MND and HC) according to the MMSE, we found interesting relationships (Table 5). The peak velocity increases, whereas kurtosis, skewness and deceleration time decrease as the MMSE score increase. Practically, the more a person is cognitively healthy intact, the faster (and decelerate for less time) and less variable the movements they make.

This work presents a pilot study, where SensRing, a novel wearable ring-shaped device, is proposed to analyse how motor performances vary during sequences of reach-to-grasp and after-grasp actions with different end-goals in healthy people and subjects with a mild decline in executive functions. Then, we investigate how these variations in motor performances can correlate with the cognitive decline. The accuracy of SensRing in measuring motor parameters in healthy subjects while performing standard tasks compared to a gold standard optoelectronic system has been already presented in [23]. In this work, we propose to use SensRing as an alternative, easy-to-use, small, wearable, non-invasive solution for clinical application in people affected by MND with deficits in EFs. Such a system, indeed, can allow objectifying the patients’ evaluation accurately, measuring their motor performances during a simple motor protocol. Shorts RG and AG sequences with different end-goals were proposed to investigate whether the different intention can influence the planning and the action both in a group of HC and d-MND subjects. The proposed experimental protocol is reliable, easy to be set-up and quick to be performed. Furthermore, it is not constrained to language or cultural issues because it is based on simple motor sequences.

The analysis of the accelerations and the angular velocities acquired by SensRing allows calculating a set of 21 motor parameters that objectively characterize the kinematic of the subjects motor performance.

We observed interesting patterns in most of the parameters when d-MND and HC subjects have been compared. These results endorse our hypothesis that EFs impairments reflect on worsening in motor performances. d-MND, indeed, needs more time to initiate the movement. Additionally, they move slower, also presenting longer deceleration phases. This suggests that a higher effort required to d-MND to perform the same tasks of HC, even in a basic motor task. Increase of movement time spent decelerating, indeed, is associated with an onward action, that requires a greater level of precision [30]. More gradual acceleration changes and longer deceleration phases are reflected also in the measure of jerk, which is the derivative of the acceleration and results slower in d-MND respect to HC. Differently, HC have lower IAV than d-MND, confirming they spend less energy to carry out the movement that results in a more optimal and ecological performance. Also, higher kurtosis and skewness values for d-MND demonstrate higher variability of pathological subjects in performing the movements. An exception to this pattern is the PASS condition for healthy that could show higher variability when approaching a partner, differently from d-MND that less modulate the kinematic of the movement according to the end-goal of the action. Finally, HC anticipate the maximum hand opening with respect to d-MND, supporting the theory that action planning in subjects with EFs disorders is compromised and they require more time to organize the grasping action, independently from the forthcoming action. Furthermore, trends found for RG phase are mainly confirmed also in AG phase, strengthen the choice of these parameters for characterizing the motor pattern of the involved subjects.

These preliminary results suggest the potentiality of this device in developing a decision support tool for clinical assessment of d-MND people, providing reliable accurate motor measurements of the subjects’ performance while carrying out a simple fast protocol. Enlarging the dataset, in future works, we trust that many of the parameter trends find out in this study can become significant differences.

Regarding the kinematic variations as response to the different end-goal of the tasks, as suggested by [30], [31], we have found that healthy subjects differently modulate the movements in the reach-to-grasp sequences according to different intentions. Conversely, people presenting EFs deficits do not show the same capability, resulting in a very low number of significant parameters able to distinguish among the three conditions (drink, place, and pass); thus they seem to adjust less their movement based on the forthcoming action. However, both HC and d-MND confirmed that more precise tasks require more effort and execution time, according to [30], [32]. Also, the deceleration phase is longer when for onward actions have higher-precision requirements. Probably, drinking is a more automatic movement, compared to place the can on a specific narrow target or to pass it to another person. Therefore, the kinematic of the first action is quite different from the other two. However, HC compared to d-MND can differentiate their performance also between the placing and the passing actions, demonstrating a finer ability in modulating kinematic according to the specific task end-goal.

Finally, the correlation of several features to the clinical score of a standard neuropsychological test (i.e., MMSE) typically used as a screening tool for cognitive assessment, confirm the reliability of these kinematic parameters in characterizing the cognitive decline (i.e., better movements are performed by people without cognitive decline).

The difference in age between the two groups is a limitation of this work. Ageing, indeed, can cause a worsening in motor performance compared to a younger group [10], [11]. However, according to [12], using goal-directed tasks, such as our experimental protocol, differences due to ageing are limited. Nevertheless, future studies should involve a larger dataset, including age-matched groups to highlight differences related to the cognitive decline. Additionally, further investigations could be performed considering the combined execution of motor planning and memory tasks [33], exploiting the paradigm of the motor and cognitive dual-task to provide insights about the cognitive resources distribution in patients with dysexecutive syndrome.

In this pilot study, SensRing has been employed to analyze the motor performances of healthy controls and subjects suffering from mild cognitive impaired, particularly a decline in executive functions, highlighted during a simple fast motor protocol based on reach-to-grasp and after-grasp sequences. The cognitively impaired subjects showed deficits in motor performance (e.g. slower movement, slower reaction, longer deceleration phases) compared to healthy subjects. Furthermore, they were less prone to modulate the kinematic of their actions during the reach-to-grasp phase (which is identical for all the tasks) when we compared similar tasks with different end-goal. These results endorse the thesis that such subjects might show impairments in action planning. Therefore, we propose to use SensRing to objectify the clinical evaluation of people suffering from mild cognitive decline, implementing a simple motor test easy to set up. Such a system could be suitable for clinical application, and then it could be also usable at home, for remote monitoring and eventually for analyzing specific tasks for neuromotor rehabilitation in patients affected by MND.

DRINK = the experimental condition to drink

PASS = the condition to pass the can to a partner

PLACE = the condition to place the can on a target

EFs = executive functions

d-MND = MND subjects with impairment in EFs, also named dysexecutive MND

MND = mild neurocognitive disorder

HC = healthy controls

RG = Reach-to-grasp action

AG = After-grasp action

MMSE = Mini Mental Score Examination

Ethics approval and consent to participate: The study was performed in accordance with the Declaration of Helsinki and approved by the ethical committee CPP Ile de France (N° IDRCB: 2019-A00342-55). All participants received detailed written explanations on the study and signed written informed consent.
Consent for publication: Not applicable.
Availability of data and materials: The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Competing interests: The authors declare that they have no competing interests.
Funding: This work is financially supported by the OLIMPIA project funded by Tuscany Region (Bando Ricerca Salute 2018, CUP J44I20000760009). This work was also supported by the following grant: ‘APAT-MOT’ grant from the FRIS (Fédération de Recherche Interventions en Santé) of the Université Cote d’Azur, Nice. Additional supports came from the Association IA, and the JL Noisiez Foundation.
Authors' contributions: ER drafted and revised the work, interpreted the data; GG acquired and analysed the data, VM participated in study conception and design, and revised the work, GM interpreted the data and revised the work, LF participated in study conception and revised the work, AG participated in the design, experimental protocol and data acquisition, PR supervised and revised the work, FC participated in study conception, supervised and revised the work. All authors read and approved the final manuscript.
Acknowledgements: Thanks to Radia Zeghari, Raphael Zory, Xavier Corveleyn and Alexandra Plonka for their support in the study development and data collection.

I. J. Deary et al., “Age-associated cognitive decline,” Br. Med. Bull., vol. 92, no. 1, pp. 135–152, 2009, doi: 10.1093/bmb/ldp033.
M. Prince, A. Comas-Herrera, M. Knapp, M. Guerchet, and M. Karagiannidou, “World Alzheimer Report 2016 Improving healthcare for people living with dementia. Coverage, Quality and costs now and in the future,” Alzheimer’s Dis. Int., pp. 1–140, 2016, [Online]. Available: https://www.alz.co.uk/research/world-report-2016.
A. D. Baddeley, “The central executive: A concept and some misconceptions,” Explor. Work. Mem. Sel. Work. Alan Baddeley, pp. 247–252, 2017, doi: 10.4324/9781315111261.
O. Godefroy, P. Azouvi, P. Robert, M. Roussel, D. Legall, and T. Meulemans, “Dysexecutive syndrome: Diagnostic criteria and validation study,” Ann. Neurol., vol. 68, no. 6, pp. 855–864, 2010, doi: 10.1002/ana.22117.
G. Dirnberger and M. Jahanshahi, “Executive dysfunction in Parkinson’s disease: A review,” J. Neuropsychol., vol. 7, no. 2, pp. 193–224, 2013, doi: 10.1111/jnp.12028.
D. Howieson, “Current limitations of neuropsychological tests and assessment procedures,” Clin. Neuropsychol., vol. 33, no. 2, pp. 200–208, 2019, doi: 10.1080/13854046.2018.1552762.
C. Ansuini, L. Giosa, L. Turella, G. Altoè, and U. Castiello, “An object for an action, the same object for other actions: Effects on hand shaping,” Exp. Brain Res., vol. 185, no. 1, pp. 111–119, 2008, doi: 10.1007/s00221-007-1136-4.
C. Becchio, V. Manera, L. Sartori, A. Cavallo, and U. Castiello, “Grasping intentions: From thought experiments to empirical evidence,” Front. Hum. Neurosci., vol. 6, p. 117, 2012, doi: 10.3389/fnhum.2012.00117.
E. Straulino, T. Scaravilli, and U. Castiello, “Social intentions in Parkinson’s disease patients: A kinematic study,” Cortex, vol. 70, pp. 179–188, 2015, doi: 10.1016/j.cortex.2015.02.012.
K. M. B. Bennett and U. Castiello, “Reach to grasp: Changes with age,” Journals Gerontol., vol. 49, no. 1, pp. 1–7, 1994, doi: 10.1093/geronj/49.1.P1.
Y. Tamaru, Y. Naito, and T. Nishikawa, “Earlier and greater hand pre-shaping in the elderly: a study based on kinematic analysis of reaching movements to grasp objects,” Psychogeriatrics, vol. 17, no. 6, pp. 382–388, 2017, doi: 10.1111/psyg.12256.
A. Cicerale, E. Ambron, A. Lingnau, and R. I. Rumiati, “A kinematic analysis of age-related changes in grasping to use and grasping to move common objects,” Acta Psychol. (Amst)., vol. 151, pp. 134–142, 2014, doi: 10.1016/j.actpsy.2014.06.004.
J. L. Alberts, M. Saling, C. H. Adler, and G. E. Stelmach, “Disruptions in the reach-to-grasp actions of Parkinson’s patients,” Exp. Brain Res., vol. 134, no. 3, pp. 353–362, 2000, doi: 10.1007/s002210000468.
L. A. R. Sacrey, T. Germani, S. E. Bryson, and L. Zwaigenbaum, “Reaching and grasping in autism spectrum disorder: A review of recent literature,” Front. Neurol., vol. 5, no. January, p. 6, 2014, doi: 10.3389/fneur.2014.00006.
A. M. B. Stoit, H. T. Van Schie, D. I. E. Slaats-Willemse, and J. K. Buitelaar, “Grasping motor impairments in Autism: Not action planning but movement execution is deficient,” J. Autism Dev. Disord., vol. 43, no. 12, pp. 2793–2806, 2013, doi: 10.1007/s10803-013-1825-8.
A. Viau, A. G. Feldman, B. J. McFadyen, and M. F. Levin, “Reaching in reality and virtual reality: A comparison of movement kinematics in healthy subjects and in adults with hemiparesis,” J. Neuroeng. Rehabil., vol. 1, pp. 1–7, 2004, doi: 10.1186/1743-0003-1-11.
T. G. Stavropoulos, A. Papastergiou, L. Mpaltadoros, S. Nikolopoulos, and I. Kompatsiaris, “Iot wearable sensors and devices in elderly care: A literature review,” Sensors (Switzerland), vol. 20, no. 10, 2020, doi: 10.3390/s20102826.
A. Moschetti, L. Fiorini, D. Esposito, P. Dario, and F. Cavallo, “Toward an Unsupervised Approach for Daily Gesture Recognition in Assisted Living Applications,” IEEE Sens. J., vol. 17, no. 24, pp. 8395–8403, 2017, doi: 10.1109/JSEN.2017.2764323.
R. Varatharajan, G. Manogaran, M. K. Priyan, and R. Sundarasekar, “Wearable sensor devices for early detection of Alzheimer disease using dynamic time warping algorithm,” Cluster Comput., vol. 21, no. 1, pp. 681–690, 2018, doi: 10.1007/s10586-017-0977-2.
E. Rovini, C. Maremmani, and F. Cavallo, “A wearable system to objectify assessment of motor tasks for supporting parkinson’s disease diagnosis,” Sensors (Switzerland), vol. 20, no. 9, pp. 1–27, 2020, doi: 10.3390/s20092630.
G. Mancioppi, L. Fiorini, M. L. Critelli, E. Rovini, M. T. Sportiello, and F. Cavallo, “Evaluation of MCI Motor Performances during a Cognitive Dual Task Exercise,” 2019 IEEE 23rd Int. Symp. Consum. Technol. ISCT 2019, pp. 247–250, 2019, doi: 10.1109/ISCE.2019.8901046.
E. M. Benssassi, J. C. Gomez, L. E. Boyd, G. R. Hayes, and J. Ye, “Wearable Assistive Technologies for Autism: Opportunities and Challenges,” IEEE Pervasive Comput., vol. 17, no. 2, pp. 11–21, 2018, doi: 10.1109/MPRV.2018.022511239.
E. Rovini, G. Galperti, L. Fiorini, G. Mancioppi, V. Manera, and F. Cavallo, “SensRing , a novel wearable ring-shaped device for objective analysis of reach- to-grasp movements,” in 2020 Annual International Conference of the IEEE Engineering in Medicine and Biology (EMBC), 2020, pp. 4020–4023.
P. Werner, S. Rabinowitz, E. Klinger, A. D. Korczyn, and N. Josman, “Use of the virtual action planning supermarket for the diagnosis of mild cognitive impairment,” Dement. Geriatr. Cogn. Disord., vol. 27, no. 4, pp. 301–309, 2009, doi: 10.1159/000204915.
M. F. Folstein, S. E. Folstein, and P. R. McHugh, “‘Mini-Mental State’ A practical method for,” J. Psychiatr. Res., vol. 12, pp. 189–198, 1975, doi: 10.3744/snak.2003.40.2.021.
B. Dubois, A. Slachevsky, I. Litvan, and B. Pillon, “The FAB: A frontal assessment battery at bedside,” Neurology, vol. 55, no. 11, pp. 1621–1626, 2000, doi: 10.1212/WNL.55.11.1621.
F. Cavallo, G. Megali, S. Sinigaglia, O. Tonet, and P. Dario, “A biomechanical analysis of surgeon’ s gesture in a laparoscopic virtual scenario .,” Stud. Health Technol. Inform., vol. 119, pp. 79–84, 2006.
D. Figo, P. C. Diniz, D. R. Ferreira, and J. M. P. Cardoso, “Preprocessing techniques for context recognition from accelerometer data,” Pers. Ubiquitous Comput., vol. 14, no. 7, pp. 645–662, 2010, doi: 10.1007/s00779-010-0293-9.
C. Becchio, L. Sartori, M. Bulgheroni, and U. Castiello, “The case of Dr . Jekyll and Mr . Hyde : A kinematic study on social intention,” vol. 17, pp. 557–564, 2008, doi: 10.1016/j.concog.2007.03.003.
K. Wilmut, M. Byrne, and A. L. Barnett, “To throw or to place: Does onward intention affect how a child reaches for an object?,” Exp. Brain Res., vol. 226, no. 3, pp. 421–429, 2013, doi: 10.1007/s00221-013-3453-0.
C. Armbrüster and W. Spijkers, “Movement planning in prehension: Do intended actions influence the initial reach and grasp movement?,” Motor Control, vol. 10, no. 4, pp. 311–329, 2006, doi: 10.1123/mcj.10.4.311.
R. G. Marteniuk, C. L. MacKenzie, M. Jeannerod, S. Athenes, and C. Dugas, “Constraints on human arm movement trajectories.,” Can. J. Psychol., vol. 41, no. 3, pp. 365–378, 1987, doi: 10.1037/h0084157.
M. Weigelt, D. A. Rosenbaum, S. Huelshorst, and T. Schack, “Moving and memorizing: Motor planning modulates the recency effect in serial and free recall,” Acta Psychol. (Amst)., vol. 132, no. 1, pp. 68–79, 2009.

TABLE 1

Kinematic Parameters Extracted From Sensring. RG = Reach-to-Grasp; AG = After-Grasp
Parameter	Meaning	Phase
IAV	Integral of the magnitude of the acceleration vector: it represents a value correlated to the energy expenditure (m/s).	RG, AG
Vpeak	Amplitude of peak velocity (m/s).	RG, AG
T_Vpeak	Time of peak velocity: it is the time instant corresponding to the peak velocity (s).	RG, AG
T_exec	Execution time: it represents the time spent to perform the movement (s).	RG, AG
T_react	Reaction time: it is the time elapsed from the start acoustic input to the beginning of the movement (s).	RG
T_decel	Deceleration time: it is the duration of the deceleration phase (s).	RG, AG
Decel%	Deceleration time (%): it represents the percentage of the deceleration phase compared to the total execution time.	RG, AG
Exc_hand	Amplitude of maximum hand excursion: it indicates the angular excursion of the hand during the grasping of the object (deg).	RG
T_exc_hand	Time of maximum excursion: it is the time instant corresponding to the maximum aperture of the hand (s).	RG
Skew	Skewness: it is a measure of the asymmetry of the distribution.	RG, AG
Kurt	Kurtosis: it is a measure of the shape of the tail of the distribution	RG, AG
rmseJerk	Root mean square of the jerk, that is the rate of change of the acceleration: it represents the smoothness of the movement (m/s³)	RG, AG

TABLE 2

CLINICAL AND DEMOGRAPHICAL DATA
	HC (10)	d-MND (8)	p-value
Female (%)	60.0%	87.5%	0.1955
Age (years)	63.7 ± 9.9	75.7 ± 5.2	0.0283
Education (%) Primary Secondary Higher	0 (0.0%) 3 (30.0%) 7 (70.0%)	0 (0.0%) 4 (50.0%) 4 (50.0%)	0.3871
MMSE	30.00 ± 0.00	24.25 ± 2.87	<0.0001

TABLE 3

MEAN VALUES AND STANDARD DEVIATION (SD) OF SIGNIFICANT KINEMATIC PARAMETERS FOR HC AND D-MND AT INTER-GROUP ANALYSIS DURING RG PHASE
	DRINK		PLACE		PASS
Parameter	HC	d-MND	HC	d-MND	HC	d-MND	p
IAV	8.49 ± 1.61	9.55 ± 1.97	9.64 ± 1.52	10.65 ± 0.98	10.13 ± 1.43	10.88 ± 1.20	NS
rmseJerk	10.58 ± 6.04	8.77 ± 3.45	9.77 ± 2.78	9.88 ± 3.80	9.28 ± 3.54	8.45 ± 2.74	NS
Skew	0.38 ± 0.08	0.41 ± 0.13	0.29 ± 0.05	0.42 ± 0.13	0.43 ± 0.11	0.37 ±0.18	+
Kurt	2.29 ± 0.13	2.52 ± 0.44	2.56 ± 0.25	2.76 ± 0.27	2.67 ± 0.20	2.83 ± 0.45	NS
Vpeak	1.17 ± 0.54	0.78 ± 0.32	1.33 ± 0.38	1.02 ± 0.35	1.13 ± 0.22	0.84 ± 0.28	+, °
T_Vpeak	0.36 ± 0.04	0.35 ± 0.03	0.36 ± 0.05	0.35 ± 0.05	0.35 ± 0.03	0.34 ± 0.05	NS
T_exec	0.88 ± 0.16	0.99 ± 0.20	0.99 ± 0.15	1.10 ± 0.10	1.04 ± 0.14	1.13 ± 0.12	NS
T_react	0.71 ± 0.11	0.77 ± 0.10	0.67 ± 0.10	0.81 ± 0.14	0.67 ± 0.10	0.77 ± 0.10	+
Hand_open_max	65.39 ± 15.83	55.99 ± 10.27	60.85 ± 11.36	51.41 ± 6.32	61.77 ±18.54	56.77 ± 9.48	NS
T_hand_open_max	0.69 ± 0.14	0.74 ± 0.10	0.63 ±0.11	0.68 ± 0.06	0.68 ± 0.19	0.69 ± 0.07	NS
T_decel	0.50 ± 0.14	0.70 ± 0.27	0.62 ± 0.13	0.75 ± 0.10	0.61 ± 0.16	0.76 ± 0.11	°
T_decel_perc	56.64 ± 4.89	63.58 ± 8.01	62.42 ± 4.21	68.32 ± 4.48	65.54 ± 2.86	68.18 ± 6.20	*, +

NS: no significance; *: Drink_HC vs Drink_d-MND; +: Place_HC vs Place_d-MND; °: Pass_HC vs Pass_d-MND

TABLE 4

MEAN VALUES AND STANDARD DEVIATION (SD) OF SIGNIFICANT KINEMATIC PARAMETERS FOR HC AND D-MND AT INTER-GROUP ANALYSIS DURING AG PHASE
	DRINK		PLACE		PASS
*Parameter*	HC	d-MND	HC	d-MND	HC	d-MND
IAV	14.56 ± 5.98	19.61 ± 9.08	13.18 ± 2.58	14.70 ± 3.85	12.39 ± 3.54	15.00 ± 5.91	NS
rmseJerk	8.62 ± 4.39	5.29 ± 2.84	6.08 ± 1.89	5.38 ± 1.38	4.36 ± 1.83	3.64 ± 0.68	*
Skew	0.34 ± 0.18	0.57 ± 0.24	0.35 ± 0.14	0.39 ± 0.12	0.23 ± 0.08	0.45 ± 0.24	*, °
Kurt	2.22 ± 0.40	2.94 ± 1.01	2.57 ± 0.31	3.01 ± 0.55	2.07 ± 0.23	2.68 ± 0.38	°
Vpeak	2.92 ± 0.94	1.73 ± 0.81	1.26 ± 0.20	0.95 ±0.32	0.73 ± 0.23	0.60 ± 0.16	*,+
T_Vpeak	0.63 ± 0.13	0.68 ± 0.20	0.35 ± 0.07	0.32 ± 0.07	0.32 ± 0.08	0.31 ± 0.08	NS
T_exec	1.51 ± 0.64	2.04 ± 0.94	1.36 ± 0.27	1.53 ± 0.39	1.27 ± 0.27	1.55 ± 0.61	NS
T_decel	0.75 ± 0.35	1.28 ± 0.65	0.99 ± 0.21	1.18 ± 0.30	0.88 ± 0.19	1.13 ± 0.44	*
T_decel_perc	50.49 ± 6.36	63.01 ± 0.79	73.69 ± 4.14	77.31 ± 2.77	73.03 ± 10.02	74.58 ± 5.62	*

NS: no significance; * significance for: DRINK_HC vs DRINK_d-MND; +: PLACE_HC vs PLACE_d-MND; * PASS_HC vs PASS_d-MND

TABLE 5

CORRELATION ANALYSIS TO MMSE SCORE
Condition	Parameter	Phase	Rho	p-value
DRINK	Skew	AG	-0.612	0.007
DRINK	Vpeak	AG	0.585	0.011
DRINK	T_decel	AG	-0.477	0.045
PLACE	Skew	RG	-0.522	0.026
PLACE	Kurt	AG	-0.491	0.039
PASS	Kurt	AG	-0.636	0.005
PASS	Vpeak	RG	0.492	0.038

Download PDF

Journal Publication

published 27 Jul, 2021

Read the published version in Journal of NeuroEngineering and Rehabilitation →

Editorial decision: Major revision
22 Feb, 2021
Review #3 received at journal
08 Feb, 2021
Review #1 received at journal
22 Jan, 2021
Review #2 received at journal
22 Jan, 2021
Reviewer #3 agreed at journal
21 Jan, 2021
Reviewer #2 agreed at journal
19 Jan, 2021
Reviewers invited by journal
18 Jan, 2021
Reviewer #1 agreed at journal
18 Jan, 2021
First submitted to journal
24 Nov, 2020
Editor assigned by journal
24 Nov, 2020
Submission checks completed at journal
24 Nov, 2020
Editor invited by journal
24 Nov, 2020

You are reading this latest preprint version

"A Wearable Ring-shaped Inertial System to Identify Action Planning Impairments During Reach-to-grasp Sequences: A Pilot Study"

Status:

Journal Publication

Version 1

Abstract

Figures

I. Introduction

Ii. Related Works

Iii. Materials And Methods

Iv. Results

V. Discussion

Vi. Conclusion

List Of Abbreviations

Declarations

References

Tables

Status:

Journal Publication

Version 1