Goals and research questions
The first goal of MUSICPLAST is to explore in-depth the development of audiovisual perception
in the healthy human brain describing in detail the corresponding cortical networks involved.
High density EEG measurements (128 channels) are going to be employed for the realization of
this goal following a stimulation protocol which incorporates auditory, visual and
audiovisual correspondences in a unified multisensory oddball paradigm. Behavioral and
neuropsychological evaluation of the multisensory perceptual abilities of the subjects will
be performed in order to achieve a holistic understanding of the phenomenon. The within
subjects comparison of the different conditions will reveal the corresponding network, while
the between subjects, its developmental trajectory. The hypothesis of this experiment, is
that the network underlying audiovisual perception will be comprised from cortical areas
connecting widespread regions including dorsolateral regions of the frontal cortex. Regarding
its development, and based on the fact that behaviorally multisensory integration continues
to improve in healthy aging, it is hypothesized that the network will be reorganized showing
enhanced density and efficiency as the age progresses.
The second goal of MUSICPLAST is to explore in-depth the plasticity of the cortical network
underlying audiovisual perception in the healthy human brain, as induced via a computer
provided short-term musical reading training. Pre- and post- training high-density EEG,
behavioral and neuropsychological measurements are going to be employed for the realization
of this goal following the same stimulation paradigm as for the previously mentioned goal.
The hypothesis of this experiment, is that the network underlying audiovisual perception will
be reorganized showing enhanced global density and efficiency, greater connectivity between
temporal and frontopolar regions and a gradual decline of the amount of plasticity as a
result of the age increase.
The third goal of MUSICPLAST is to evaluate the effectiveness of the proposed computerized
music reading training protocol as a tool that will promote beneficial neuroplastic changes
in pathological populations and specifically, dyslectic adolescents and elderlies. Pre- and
post- training high-density EEG, behavioral and neuropsychological measurements are going to
be employed for the realization of this goal following the same stimulation paradigm as for
the previously mentioned goals. The hypothesis of this experiment, is that the network
underlying audiovisual perception will be beneficially reorganized due to the training
showing enhanced global density and efficiency, greater connectivity between temporal and
frontopolar regions and that these changes will be accompanied from improvement in the
corresponding behavioral and neuropsychological evaluation.
Research design and methodology Participants
A sample size analysis was conducted via G-Power based on the data from a previous relevant
study of the PI and indicated that the desired sample size in order to reach a power of 0.8
and a two sided significance level of a = 0.05 is 13 subjects in each group. In order to
account for possible dropouts, MUSICPLAST will include 20 subjects in each group [adolescents
(n = 20), young adults (n = 20), elderlies (n = 20) and controls (n=20)]. The study will be
conducted according to the declaration of Helsinki with regard to the participation of human
subjects in research.
EEG and behavioral measurements
Prior to the beginning of the training and after its completion, the participants will be
subjected into 2 electroencephalographic measurements. The recordings will be performed in an
electromagnetically shielded room located in the Medical Physics Laboratory of AUTH using a
Nihon Kohden EEG-1200 system with 128 active electrodes. The first measurement will be of
resting state activity, allowing a network analysis of resting state cortical connectivity
while the second will record Event Potentials via a multisensory oddball paradigm which is
appropriately adjusted for audiovisual perception, as it incorporates audiovisual, auditory
and visual mismatches within one run. This paradigm has been proposed and successfully
applied in a series of studies by the PI and it allows both a spatiotemporal analysis of the
cortical activity as well as a functional connectivity analysis.
Concurrently to the EEG measurements the subjects will respond behaviorally via button
presses to an audiovisual congruency task on which they have to evaluate whether the images
presented correspond to the tones they hear based on the rule "the higher the pitch of the
tone - the higher the position of the circle". This measurement will allow the behavioral
evaluation of the discrimination accuracy of congruent and incongruent audiovisual stimuli.
Prior to the beginning of the training and after its completion, the participants will be
subjected into a neuropsychological evaluation, which will consist of a battery including the
following tests:
All subjects:
Stroop Test Trail A and B Digit Span Test Performance Subtest of Wechsler Intelligence Scale
Children III or Wechsler Adult Intelligence Scale according to the age.
Elderlies only:
Mini Mental State Examination (MMSE) Montreal Cognitive Assessment (MoCA) Geriatric
Depression Scale
Dyslectics only:
Word and nonword reading Word and nonword spelling Phonological ability
Intervention - MUSICPLAST training protocol
Each subject will receive 20 sessions of the computerized training based on a simplified
music reading within a period of 4 weeks. The training will be provided via a smartphone or a
tablet. The application will record a log file for each training session, which will last for
20 minutes and the log file will be directly available to the researcher, allowing him/her to
follow the timely participation and commitment of the subject. The stimuli used in the
training protocol will be constructed along the same principles of the audiovisual stimulus
patterns used in the EEG recordings: a visual image will show a circle on top of a background
of 5 horizontal lines. The duration of the patterns will be of a mean duration of 10 seconds.
The position of the circle within the five lines will either follow the rule "the higher the
pitch of the tone, the higher the position of the circle" or it will violate the above rule.
This rule was in principle followed also in previous studies conducted by the PI with
positive results regarding the induced neuroplasticity. The visual images will only represent
the frequency of the pitch and not the rhythmical pattern. The participant will have to
respond at the end of each trial via a button press whether the audiovisual input confirmed
or violated the above presented rule. A feedback image of a smiling face for the correct
responses or a not smiling face for the incorrect responses will be provided for each trial.
The protocol will be of adjusted difficulty, having 5 levels of difficulty which will be
based on 2 different factors: the presentation speed, and the difficulty of the musical
pattern. In contrast to the EEG measurement procedure, during which the melodic patterns will
be unknown to the subjects, the melodic patterns used in the training will be comprised of
short but identifiable parts of known musical pieces in order to engage the participants and
to provoke emotional response. The training will be improved in comparison to the
corresponding previous studies of the PI by providing feedback and by the use of known
musical pieces which will induce an emotional response to the trainees.
EEG data analysis and statistical procedures
The EEG data collected will be pre-processed using the Brain Electrical Source Analysis
software (BESA research, version 6, Megis Software, Heidelberg, Germany). The recorded data
will be separated in epochs of 1000 ms including a pre-stimulus interval of 200 ms. Epochs
will be baseline corrected using the interval from -100 to 0 ms. Current Density
Reconstructions (CDRs) will be calculated on the neural responses of each subject and each
condition using the sLORETA method55 as provided by BESA. The Matlab ® (The MathWorks, Inc.,
Natick, Massachusetts, United States.) toolbox HERMES56 will be used for calculating the
adjacency matrix from the voxel time-series. The toolbox Network Based Statistic (NBS) will
be used to statistically identify significant connections in the graphs, using a General
Linear Model approach. For the statistical analysis of the data a 2 × 2 × 2 × 2 mixed model
ANOVA design will be followed, with 3 within subjects factors: condition (congruent and
incongruent), modality (audiovisual, auditory, and visual) and time (pre- and post- training)
and one between subjects: (age group). The significance level will be set to p < 0.05
corrected for multiple comparisons via Family Wise Error correction.