Event-Related Desynchronization Analysis During Action

Observation and Motor Imagery of Transitive Movements

Stefania Coelli

, Alessandra Calcagno

, Federico Temporiti

1,2 c

, Roberto Gatti

2,3 d

Manuela Galli

and Anna M. Bianchi

Department of Electronic, Information and Bioengineering, Politecnico di Milano, Milan, Italy

Physiotherapy Unit, Humanitas Clinical and Research Center - IRCCS, Rozzano, Milan, Italy

Department of Biomedical Sciences, Humanitas University, Pieve Emanuele, Milan, Italy

roberto.gatti@hunimed.eu

Keywords: Electroencephalography, Action Observation, Motor Imagery.

Abstract: Rehabilitation and motor skill learning approaches based on Action Observation (AO) and Motor Imagery

(MI) rely on the assumption that the sensorimotor system is stimulated by AO and MI tasks similarly to the

actual execution of a movement. An advantage of AO over MI is that it is less dependent on subject’s

imagination ability, and a direct comparison of their effect on cortical activations during complex upper limb

movements has been rarely examined. Therefore, in this study we compare sensorimotor event related

desynchronization (ERD) patterns, as a measure of cortical activation, collected from 46 healthy volunteers

performing AO and MI protocols. In both mu and beta sensorimotor rhythms a stronger ERD was elicited by

AO, characterized by an evident lateralization in the contralateral side of the brain with respect to the limb

involved in the observed movement.

1 INTRODUCTION

The mirror mechanism is related to the response of

the brain that transforms the visual perception of

actions, performed by others, into a motor

representation in the brain of the observer (Rizzolatti

and Sinigaglia, 2010). It has been shown that, during

action observation (AO), the cortical areas that are

normally activated during motor execution (ME), are

similarly activated, supporting the existence of the so-

called motor-resonance phenomenon, even if several

factors influence the patterns and the strength of such

response (Kemmerer, 2021). These factors may be

grouped in four categories according to Kemmerer: i)

relation between agent and observer, ii) factors

involving the action, iii) factors involving the actors

and iv) factors related to the observer. Also the

action’s context may play a role (Kemmerer, 2021).

https://orcid.org/0000-0003-4104-4790

https://orcid.org/0000-0003-3989-1712

https://orcid.org/0000-0003-1771-2365

https://orcid.org/0000-0002-4669-1287

https://orcid.org/0000-0003-2772-4837

https://orcid.org/0000-0002-8290-7460

Researches in recent years have demonstrated, for

example, that the observation of a movement from a

first person perspective produces a stronger

modulation of the Rolandic sensorimotor rhythms

(Angelini et al., 2018; Drew et al., 2015). Moreover,

watching a transitive motor task (i.e., object directed)

is more engaging than observing an intransitive action

(Coll et al., 2017). The possibility of using the mirror

mechanism to stimulate the sensorimotor system has

been exploited as an innovative rehabilitation

approach called Action Observation Therapy

(Calcagno et al., 2022; Rizzolatti et al., 2021;

Temporiti et al., 2020). This motor learning approach

could be adopted to facilitate another promising

framework for the acquisition and recovery of motor

skills, that is the internal simulation of motor action,

i.e. Motor Imagery (MI) (Daeglau et al., 2021;

Gonzalez-Rosa et al., 2015). Like AO, MI has been

Coelli, S., Calcagno, A., Temporiti, F., Gatti, R., Galli, M. and Bianchi, A.

Event-Related Desynchronization Analysis During Action Observation and Motor Imagery of Transitive Movements.

DOI: 10.5220/0011740600003414

In Proceedings of the 16th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2023) - Volume 4: BIOSIGNALS, pages 86-93

ISBN: 978-989-758-631-6; ISSN: 2184-4305

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

shown to activate the similar brain network also

supporting ME in studies employing

Electroencephalography (EEG) to monitor brain

response (Gonzalez-Rosa et al., 2015; Neuper et al.,

2005). Nevertheless, in term of rehabilitation

practice, the efficacy of MI paradigms is limited by

the ability of the subjects in performing a correct

imagination task, even of simple movements, while

the observation of complete and transitive

movements is believed to strongly activate the

sensorimotor cortex.

The dynamical activation of the brain during

motor- and sensorial- related stimulation is typically

measured by the event-related desynchronization

(ERD) of the sensorimotor cortical rhythms in the mu

(8-12 Hz) and beta (14-24 Hz) EEG frequency bands

over central motor areas of the brain (Neuper et al.,

2005; Tacchino et al., 2017). Few studies presented a

direct comparison of the ERD patterns characterizing

AO and MI of complex movement (Gonzalez-Rosa et

al., 2015), in order to understand to which extent they

overlap. To address this topic, in the current study, we

compare sensorimotor ERD patterns extracted from

EEG signals acquired on a group of healthy young

volunteers performing AO and MI protocols.

Specifically, three complex transitive manual

dexterity tasks were employed and the effect of the

different complexity of the movement was further

studied.

2 MATERIALS AND METHODS

2.1 Data Acquisition

During the experiment, EEG signals were collected

from 46 right-handed healthy participants (Age: 20-

30, 22 female) using a 61-channel cap and the SD

LTM 64 express polygraph recording system

(Micromed, Mogliano Veneto, Italy). Signals were

sampled at 1024 Hz and the impedances were kept

under 20KOhm using a conductive hydrogel.

The action observation (AO) and motor imagery

(MI) protocol was approved by the Internal Ethical

Committee of the Istituto Clinico Humanitas

(Rozzano, Italy). All the subjects signed an informed

consent before the recordings. The stimulation

sequence consisted in the presentation of a 6.5-s-long

video-clip containing an upper limb movement

performed from the visual perspective of the subject

(1st person) and executed by a gender-matched actor.

Only the upper limb of the actor was visible. The

video-clip was preceded by a 3-s period of rest

(fixation of a cross) and 2 seconds of preparation (red

dot) displayed on a screen positioned in front of the

participant. The stimulation sequence was repeated

for 20 trials. The same sequence was repeated for the

motor imagery task but, in this case, only the first

frame of the video was shown for the same amount of

time (6.5 seconds). During the motor imagery task,

participants were asked to image performing the

movement themselves. Again, 20 trials were

recorded. AO and MI sequences together formed a

single stimulation block. Three stimulation blocks

(W1, W2 and W3) were delivered to participants

separated by resting periods during which volunteers

were free to move. In each block a different transitive

movement was shown in the video-clip (Figure 1).

The three movements were characterized by a

different level of interaction with objects. W1

consisted in picking-up five small coins, W2

presented the use of a hammer to hit a nail, and W3

displayed the interaction with tweezers to move a

small object into a plastic glass. The presentation

order of the videos was randomized.

2.2 Data Pre-Processing and Analysis

EEG signals were pre-processed using EEGLAB

toolbox and custom scripts optimized for the study

aim (Cassani et al., 2022). First, data were band-pass

filtered between 1 and 45 Hz with a FIR, zero-phase

filter, down sampled to 256 Hz and bad channels were

visually selected and removed. Signals were cut into

epochs from -5 to +6.5 seconds with respect to the

main stimulus presentation (start of the video/frame

presentation). The extended Infomax independent

component analysis was applied to the concatenated

epochs and with the support of the IClabel plugin

(Pion-Tonachini et al., 2019), the source of artifacts

were identified and removed. The previously rejected

bad channels were interpolated, and signals were re-

referenced to the common average reference. Finally,

epochs with residual artefacts were visually checked

and rejected.

Cleaned trials of each participant, separately for

AO and MI were used to compute the time-frequency

representation. The time-frequency analysis was

performed through EEGLAB toolbox using Morlet

wavelets starting from 3 cycle and expanding linearly

with the frequency for continuous transform as

suggested in the literature (Angelini et al., 2018;

Avanzini et al., 2012; Tacchino et al., 2017). EEG

power values were calculated for 145 linearly spaced

frequencies (from 4 Hz to 30 Hz) and along 200 time

bins resulting in a time resolution of ~0.05 seconds.

To select both the individual baseline period and the

mu frequency range, the two-second period from -4

Event-Related Desynchronization Analysis During Action Observation and Motor Imagery of Transitive Movements

Figure 1: Stimulation sequence for action observation (AO) and motor imagery (MI) tasks.

to -2 sec with respect the main stimulus presentation,

corresponding to the cross fixation, was analyzed at

C3 channel position.

We identified the best baseline interval as the 1-s-

long segment (50% overlapping moving window)

showing the highest power value associated to the

averaged alpha power between 8 and 12 Hz. Once the

baseline had been selected, the individual mu

frequency (IMF) was identified as the power peak

between 8 and 12 Hz in the baseline. This procedure

was repeated for the six conditions (AO/MI; three

videos W1, W2 and W3). While the specific baseline

was selected in each condition, the final IMF was

obtained as the median of the six values extracted. We

then defined the mu band as the frequency range

between IMF-1 Hz and IMF+1 Hz, while the standard

low beta frequency range was used [14 - 20] Hz

(Angelini et al., 2018). In the two frequency ranges

the %ERD was finally computed along each time bin

t as in (1)

%ERD(t)=(P(t)-B)/B*100 (1)

where P(t) is the mean power in the analyzed

frequency range at each time-point, and B the power

of the same frequency range averaged in the selected

baseline period. The ERD time course was divided

into consecutive and not overlapping 1-s-long time

windows from -1 to +4 seconds and the mean %ERD

value for each window was computed. The ERD time

course was further analyzed restricted to six brain

regions of interest (B-ROI) averaging the ERD at the

channel position associated to each B-ROI: Frontal

left (FRL: F3, FC1, FC3), Frontal right (FRR: F4,

FC2,FC4 ), Central Left (CL: C1, C3, C5), Central

right (CR: C2, C4, C6), Centro-Parietal left (CPL:

CP3, CP1, P3) and Centro-Parietal right (CPR: CP4,

CP6 , P4). The asymmetry of the Centro-parietal area

is due to the removal of some EEG channels (e.g.,

CP5 and CP2) operated by the acquisition system in

order to simultaneously acquire EMG bipolar signals.

Two repeated measure ANOVAs were applied to

the data, one for each frequency range, with 4 within

factors: 3 video Types (W) x 6 B-ROIs x five Time

windows (f0 = pre stimulus ERD, from f1 to f4 post

stimulus segments) x 2 Tasks (AO and MI). Data

were first tested for normality and log-transformed

when necessary. Outliers were also detected (> 3*SD

of the ERD percentage value) and if the participants

were identified as outlier in at least two windows and

more than two brain areas, their data were discarded.

This choice was made to easily and automatically

recognize subjects with an abnormal behaviour (6

subject were removed). When the sphericity of the

variances was not respected, the Greenhouse-Geisser

correction was applied.

3 RESULTS

3.1 Individual Mu-Rhythm

Modulations

Figure 2 displays for each time window and video

type, the mean and standard error values for both

action observation and motor imagery tasks

computed on the final set of 40 participants.

The ANOVA test identified a significant main effect

for the factor Task (F(1,40)=20.39; p= 5.46e-05), B-

ROIs (F(3.4,136.2)=22.83; p = 4.73e-13) and Time

(F(1.5,61.2)=37.4; p= 7.27e-10), but not for video

Type (F(2,80)=0.69; p= 0.5)). A significant Task*B-

ROI*Time interaction was detected (F(8.5,

341.2)=2.04; p=0.037), and the following two-factors

interactions: Task*B-ROI (F(3.65, 145.9)=6.73; p=

BIOSIGNALS 2023 - 16th International Conference on Bio-inspired Systems and Signal Processing

Figure 2: Mean and standard error of the mean (SEM) of mu-rhythm %ERD values for the five windows of interest (-1s to

+4 s) of both AO and MI tasks in each analyzed region of the scalp. Colours represent different video types. CL: central left,

CPL: centro-parietal left, CPR: centro parietal right, CR: central right, FL: frontal left and FR: frontal right.

9.45e-05), B-ROI*Time (F(5.8,232.8)=13.5; p= 7e-

13), Task*Time (F(2.43,97.4)=4.2; p=0.013). Finally

the interaction W*Time was also significant

(F(5.7,226.6)=2.27; p= 0.042).

Splitting by brain regions, significant interaction

between task and time were found in three regions,

namely CL (F(3.1,377.6)=5.3; p=0.001),

CPL(F(3.1,381.8)=9.4; p=3.42e-06) and

CPR(F(3.2,390.4)=9.4; p= 0.005). In these regions, the

task effect was significant in each window (p<0.005)

indicating a stronger ERD for the AO task (Table 1).

Splitting the ANOVA by tasks, we found for the

AO a significant interaction B-ROI*Time (F(6.99,

286.51)=0.349, p= 2.64e-11).

Table 1: P-values of the significant differences between

task type (AO vs MI) in each B-ROIs and window of the

mu ERD. CL: central left, CPL: centro-parietal left, CPR:

centro parietal right, CR: central right, FL: frontal left and

FR: frontal right.

ROI f0 f1 f2 f3 f4

CL 0.001 9.6e-10 0.0001 8.8e-06 3.1e-06

CPL 0.005 4.5e-10 0.0002 0.0001 9.2e-07

CPR 0.003 7.3e-09 7.2e-05 4.5e-05 1.1e-06

CR 0.004 6.2e-07 0.0002 0.001 3.2e-05

FRL 0.002 0.02 0.002

FRR 0.018 3.5e-05 0.008 0.004

Splitting again by time, in every time window after

the stimulus presentation a significant effect of the

ROI was found, while no effect was found in f0 after

Bonferroni's correction. The post-hoc analysis with

Bonferroni’s correction (Table 1) showed a stronger

mu ERD in the left centro-parietal (CPL) ROI with

respect to all the other ROIs in each time window

during video observation. The left central ROI

showed a significantly stronger ERD in each window

with respect to CR, supporting the lateralization of the

mu-rhythm modulation. The two frontal areas were

never different. Investigating the time effect in each

B-ROI (Table 2) we found that in each region, f0 was

different from all the other time windows suggesting

a strong effect of the video presentation (p <0.0001).

Moreover, significant differences were observed in

CL, CPL and CR between f2 and f3, due to a partial

re-synchronization in f3 (p

=0.025; p

CPL

=0.001 and

=0.033).

Concerning the MI, we found a significant B-

ROI*Time interaction (F(6.29, 251.58)=0.4, p=

4.28e-08), but also a W*time significant interaction

(F(6.17, 246.6)=0.77, p =0.02). Exploring the first, in

every time window a significant effect of the ROI was

found. The post-hoc analysis with Bonferroni's

correction showed a stronger mu ERD in the left

centro-parietal (CPL) ROI with respect to all the other

ROIs in each time window from f1 to f4, but also with

respect to CL, CPR and CR in f0.

Event-Related Desynchronization Analysis During Action Observation and Motor Imagery of Transitive Movements

Table 2: Corrected p-values of the significant differences among B-ROIs in each task and window of the mu ERD. CL: central

left, CPL: centro-parietal left, CPR: centro parietal right, CR: central right, FL: frontal left and FR: frontal right.

AO MI

f0 f1 f2 f3 f4 f0 f1 f2 f3 f4

CL Vs CPL 1.0e-13 1.6e-12 9.3e-14 1.0e-11 0.042 1.7e-13 3.1e-13 7.1e-10 2.2e-09

CL Vs CPR

CL Vs CR 1.2e-08 1.9e-10 3.96e-11 4.86e-07 8.4e-05 7.0e-07 1.1e-05 0.0002

CL Vs FRL 0.045 0.01

CL Vs FRR

CPL Vs CPR 1.4e-07 2.2e-10 3.7e-10 1.64e-06 0.035 1.8e-08 6.1e-11 4.3e-08 5.2e-07

CPL Vs CR 4.5e-18 6.0e-21 5.18e-24 3.52e-16 3.88e-05 4.48e-17 7.9e-19 3.9e-14 8.8e-13

CPL Vs FRL 6.8e-14 2.1e-16 2.4e-13 1.1e-10 1.1e-06 1.3e-08 9.7e-06 2.7e-05

CPL Vs FRR 7.8e-14 1.6e-17 8.2e-16 3.3e-12 3.7e-07 3.2e-09 1.5e-05 2.5e-05

CPR Vs CR 1.1e-07 0.0009 1.2e-05 6.5e-06 0.007 0.033

CPR Vs FRL

CPR Vs FRR 0.021

CR Vs FRL 0.001 0.007 0.0007 1.3e-07 4.3e-06 6.9e-07 1.4e-07

CR Vs FRR 0.0004 0.0009 8.78e-08 0.0002 0.019 7.8e-10 6.2e-10 1.2e-09 1.8e-08

FRL Vs FRR

In all the time windows, CR was found significantly

different from both the frontal regions and the left

central and centro-parietal ones. Specifically, CR

showed a weaker ERD (Table 2). Similarly, to the AO

case, exploring the effect of the time in each B-ROI,

f0 was different from all the other time windows

suggesting an effect of the video presentation (p<

0.0001). Moreover, in CL and CPL a partial

resynchronization was observed in f3 and f4 with

respect to f2 (p<0.05).

For the MI case, we further explored the effect of

the video type focusing on each window from f1 to f4

in which the video was presented. In f1, W1 showed

the less strong ERD (p < 0.001), in f2 no differences

were significant, in f3 W1 showed an overall re-

synchronization, while W3 induced a more persistent

ERD (p<0.001).

3.2 Beta Band Modulations

Figure 3 shows for each time window and video type,

the mean and standard error values of both action

observation and motor imagery ERD.

The ANOVA test identified a significant main

effect for the factor Task (F(1,41)= 14.7, p=0.0004),

B-ROIs (F(3.44,141)=25.22, p= 2.5e-14) and Time

(F(2.02,82.9)= 58.3, p= 1.24e-16).

A significant Task*B-ROI*Time interaction was

detected (F(9.8, 402.34)= 2.21, p=0.017), and the

following two-factors interactions: Task*B-ROI

(F(3.44,141.2)=3.23, p=0.019), B-ROI*Time

(F(6.78,278)=11.1, p=4.33e-12), Task*Time

(F(3.1,125.4)=3.6, p=0.015). Nor the main effect

neither the interactions including the video Type

factor were found significant.

Splitting by brain regions, significant interaction

between task and time were found in three regions,

namely CL (F(3.7,460.7)=6.24, p=0.0001),

CPL(F(3.4,421.8)=8.1, p=1.17e-05) and

CPR(F(3.3,411.9)=2.8 , p=0.036), as for the mu band.

In these regions, the task effect was significant in

each window indicating a stronger ERD for the AO

task as reported on (Table 3).

Splitting the ANOVA by tasks, we found for the

AO a significant interaction B-ROI*Time (F(6.03,

253.34)=6.3, p = 3.33e-06). Splitting the again by

time, in every time window a significant effect of the

ROI was found (Table 4). The post-hoc analysis with

Bonferroni's correction showed a stronger mu ERD in

the left centro-parietal (CPL) ROI with respect to all

the other ROIs in each time window during video

observation and only with respect to CR and FRR

in f0.

Table 3: P-values of the significant differences between

task type (AO vs MI) in each B-ROIs and window of the

Beta ERD.

ROI f0 f1 f2 f3 f4

CL 4.9e-06 9.7e-05 0.004 0.026

CPL

1.5e-07 0.0002 0.006 0.043

CPR 0.008 2.97e-06 0.0001 0.0002 4.8e-05

CR 0.024 0.0001 0.003 0.04

FRL

0.002 0.024 0.037

FRR

0.0008 0.035

BIOSIGNALS 2023 - 16th International Conference on Bio-inspired Systems and Signal Processing

Figure 3: Mean and standard error of the mean (SEM) of beta band %ERD values for the five windows of interest (-1s to +4

s) of both AO and MI tasks in each analyzed region of the scalp. Colours represent different video types. CL: central left,

CPL: centro-parietal left, CPR: centro parietal right, CR: central right, FL: frontal left and FR: frontal right.

The left central ROI showed a significantly

stronger ERD in each window with respect to the CR

from f1 to f4, supporting the lateralization of the mu-

rhythm modulation. The two frontal areas were never

different. Investigating the time effect in each B-ROI

we found that in each region, f0 was different from all

the other time windows suggesting a strong effect of

the video presentation (p<0.0001). The windows f1 to

f4 were not different in all the B-ROI except for CPL,

where f3 and f4 showed a significant re-

synchronization with respect to f1 and f2 (p<0.05).

Concerning the MI, we found a significant B-

ROI*Time interaction (F(9.89, 405.3)=3.3, p=0.0004).

In every time window a significant effect of the ROI

was found. The post-hoc analysis with Bonferroni's

correction showed a stronger beta ERD in the left

central (CL) and left centro-parietal (CPL) ROI with

respect to all the other ROIs in each time window

during the MI task. CR and CPR were found

significantly different from both CL and CPL also

during the red-circle pre-task period (f0).

Similarly, to the AO case, exploring the effect of

the time in each B-ROI, f0 was different from all the

other time windows suggesting an effect of the MI task.

4 DISCUSSION AND

CONCLUSIONS

The aim of this work was to compare the dynamical

cortical activation patterns observed during AO and MI

tasks in a group of healthy young volunteers. As a

measure of sensorimotor response to the presented

stimulations, we computed the ERD time course in

both individual mu and standard beta frequency ranges.

To increase the stimulation effect, the video-clips

used in the experiment comprised complex upper

limb object-directed (transitive) movements

performed from the observer’s perspective (Angelini

et al., 2018; Coll et al., 2017). For the MI task, the

same movements were asked to be imagined and a

frame of the video was shown to facilitate the

imagination. In all the explored brain regions, a

stronger mu ERD was elicited by the observation,

rather than the imagination of the movement, in line

with previous study by Gonzalez-Rosa et al., 2015.

Even so, the activation patterns were similar, with an

evident lateralization over the contralateral brain

areas and, in particular, a stronger engagement of the

CPL region, which can be associated to the

somatosensory cortex.

This latter results is in line with the hypothesis

that both AO and MI brain response may be more

correctly related to the sensory integration rather than

to the actual motor execution functions (Coll et al.,

2017). Since sensorimotor mu and beta oscillations

are not completely independent (Tacchino et al.,

2020), the modulation of the beta power followed the

same trend of significance detected in the mu band.

In this direction, further investigation at the source

level would provide a more precise distinction

between the two Rolandic oscillations.

Event-Related Desynchronization Analysis During Action Observation and Motor Imagery of Transitive Movements

Table 4: Corrected p-values of the significant differences among B-ROIs in each task and windows of the beta ERD. CL:

central left, CPL: centro-parietal left, CPR: centro parietal right, CR: central right, FL: frontal left and FR: frontal right.

AO MI

f0 f1 f2 f3 f4 f0 f1 f2 f3 f4

CL Vs CPL

3.1e-12 5.6e-09 5.1e-07 2.3e-08 1.8e-07 2.3e-07 5.6e-09 1e-06

CL Vs CPR

0.03 0.026 0.023 0.002 3.7e-05 0.001

CL Vs CR

5.3e-09 4.0e-08 1.5e-08 0.002 0.0009 6.1e-07 7.5e-06 0.0006 0.031

CL Vs FRL

CL Vs FRR

0.006 0.002 0.028 0.047

CPL Vs CPR

1.8e-10 5.7e-08 1.6e-08 4.4e-05 0.0008 1.0e-09 4.5e-11 7.6e-14 1.7e-10

CPL Vs CR

0.002 2.3e-23 9.8e-18 6.3e-17 1.5e-11 9.1e-05 5.0e-15 1.8e-13 1.0e-14 1.0e-10

CPL Vs FRL

4.8e-12 7.1e-10 2.1e-06 2.7e-06 1.9e-06 8.4e-06 6.2e-07 7.6e-05

CPL Vs FRR

0.002 1.2e-15 4.9e-14 2.7e-11 1.2e-10 0.014 7.5e-10 1.16e-10 3.8e-07 1.5e-06

CPR Vs CR

CPR Vs FRL

0.003 9.0e-06 0.001

CPR Vs FRR

CR Vs FRL

0.002 0.0004 0.001 0.0007 0.002 0.006

CR Vs FRR

FRL Vs FRR

Overall, the increasing complexity of the

movement was not a significant factor, even if some

actions seem to be more difficult to imagine than

others. Interestingly, the small influence of the video

type, present for the mu rhythm modulation, was

absent for the beta band, where the ERD pattern was

more consistent across video type and cortical

regions.

In conclusion, current results support the

potentiality of an action observation approach for

stimulating the sensorimotor system with a less

reliance on the subject’s imaginative abilities,

essential for achieving good results in motor imagery

protocols. Nevertheless, further studies are needed to

test the efficacy of an AO intervention alone on motor

skills learning and its effect on brain rhythm

modulation patterns.

ACKNOWLEDGEMENTS

The authors would like to thank students Costanza

Neve, Elisa Morganti, Riccardo Matarazzo e

Francesco Latino for their help in data pre-

processing.

REFERENCES

Angelini, M., Fabbri-Destro, M., Lopomo, N.F., Gobbo,

M., Rizzolatti, G., Avanzini, P., 2018. Perspective-

dependent reactivity of sensorimotor mu rhythm in

alpha and beta ranges during action observation: an

EEG study. Sci. Rep. 8, 1–11. https://doi.org/

10.1038/s41598-018-30912-w

Avanzini, P., Fabbri-Destro, M., Dalla Volta, R., Daprati,

E., Rizzolatti, G., Cantalupo, G., 2012. The dynamics

of sensorimotor cortical oscillations during the

observation of hand movements: An EEG study. PLoS

One 7, 1–10. https://doi.org/10.1371/journal.

pone.0037534

Calcagno, A., Coelli, S., Temporiti, F., Mandaresu, S.,

Gatti, R., Galli, M., Bianchi, A.M., 2022. Action

Observation Therapy Before Sleep Hours: An EEG

Study. Annu. Int. Conf. IEEE Eng. Med. Biol. Soc.

IEEE Eng. Med. Biol. Soc. Annu. Int. Conf. z2022,

4809–4812. https://doi.org/10.1109/EMBC48229.

2022.9871733

Cassani, C.M., Coelli, S., Calcagno, A., Temporiti, F.,

Mandaresu, S., Gatti, R., Galli, M., Bianchi, A.M.,

2022. Selecting a pre-processing pipeline for the

analysis of EEG event-related rhythms modulation.

Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. IEEE Eng.

Med. Biol. Soc. Annu. Int. Conf. 2022, 4044–4047.

https://doi.org/10.1109/EMBC48229.2022.9871394

Coll, M.-P., Press, C., Hobson, H., Catmur, C., Bird, G.,

2017. Crossmodal Classification of Mu Rhythm

Activity during Action Observation and Execution

Suggests Specificity to Somatosensory Features of

Actions. J. Neurosci. 37, 5936–5947. https://doi.org/

10.1523/JNEUROSCI.3393-16.2017

Daeglau, M., Zich, C., Welzel, J., Saak, S.K., Scheffels,

J.F., Kranczioch, C., 2021. Event-related

desynchronization in motor imagery with EEG

neurofeedback in the context of declarative interference

and sleep. Neuroimage: Reports 1, 100058.

https://doi.org/10.1016/j.ynirp.2021.100058

BIOSIGNALS 2023 - 16th International Conference on Bio-inspired Systems and Signal Processing

Drew, A.R., Quandt, L.C., Marshall, P.J., 2015. Visual

influences on sensorimotor EEG responses during

observation of hand actions. Brain Res. 1597, 119–128.

https://doi.org/10.1016/j.brainres.2014.11.048

Gonzalez-Rosa, J.J., Natali, F., Tettamanti, A., Cursi, M.,

Velikova, S., Comi, G., Gatti, R., Leocani, L., 2015.

Action observation and motor imagery in performance

of complex movements: Evidence from EEG and

kinematics analysis. Behav. Brain Res. 281, 290–300.

https://doi.org/10.1016/j.bbr.2014.12.016

Kemmerer, D., 2021. What modulates the Mirror Neuron

System during action observation?: Multiple factors

involving the action, the actor, the observer, the

relationship between actor and observer, and the

context. Prog. Neurobiol. 205, 102128. https://

doi.org/10.1016/j.pneurobio.2021.102128

Neuper, C., Scherer, R., Reiner, M., Pfurtscheller, G., 2005.

Imagery of motor actions: Differential effects of

kinesthetic and visual–motor mode of imagery in

single-trial EEG. Cogn. Brain Res. 25, 668–677.

https://doi.org/10.1016/j.cogbrainres.2005.08.014

Pion-Tonachini, L., Kreutz-Delgado, K., Makeig, S., 2019.

ICLabel: An automated electroencephalographic

independent component classifier, dataset, and website.

Neuroimage 198, 181–197. https://doi.org/10.1016/j.

neuroimage.2019.05.026

Rizzolatti, G., Fabbri-Destro, M., Nuara, A., Gatti, R.,

Avanzini, P., 2021. The role of mirror mechanism in the

recovery, maintenance, and acquisition of motor

abilities. Neurosci. Biobehav. Rev. 127, 404–423.

https://doi.org/10.1016/j.neubiorev.2021.04.024

Rizzolatti, G., Sinigaglia, C., 2010. The functional role of

the parieto-frontal mirror circuit: interpretations and

misinterpretations. Nat. Rev. Neurosci. 11, 264–274.

https://doi.org/10.1038/nrn2805

Tacchino, G., Coelli, S., Reali, P., Galli, M., Bianchi, A.M.,

2020. Bicoherence Interpretation in EEG Requires

Signal to Noise Ratio Quantification: An Application to

Sensorimotor Rhythms. IEEE Trans. Biomed. Eng. 67.

https://doi.org/10.1109/TBME.2020.2969278

Tacchino, G., Gandolla, M., Coelli, S., Barbieri, R.,

Pedrocchi, A., Bianchi, A.M., 2017. EEG Analysis

during Active and Assisted Repetitive Movements:

Evidence for Differences in Neural Engagement. IEEE

Trans. Neural Syst. Rehabil. Eng. 25.

https://doi.org/10.1109/TNSRE.2016.2597157

Temporiti, F., Adamo, P., Cavalli, E., Gatti, R., 2020.

Efficacy and Characteristics of the Stimuli of Action

Observation Therapy in Subjects With Parkinson’s

Disease: A Systematic Review. Front. Neurol. 11.

https://doi.org/10.3389/fneur.2020.00808.

Event-Related Desynchronization Analysis During Action Observation and Motor Imagery of Transitive Movements