Brain, music and emotion: An EEG proof-of-concept study on musically continuous, non-personalized emotional responses

It has been repeatedly reported that motivation for listening to music is majorly driven by the latter’s emotional effect. There is a relative opposition to this approach, however, suggesting that music does not elicit true emotions. Counteracting this notion, contemporary research studies indicate that listeners do respond affectively to music providing a scientific basis in differentially approaching and registering affective responses to music as of their behavioral or biological states. Nevertheless, no studies exist that combine the behavioral and neuroscientific research domains, offering a cross-referenced neuropsychological outcome, based on a non-personalized approach specifically using a continuous response methodology with ecologically valid musical stimuli for both research domains. Our study, trying to fill this void for the first time, discusses a relevant proof-of-concept protocol, and presents the technical outline on how to multimodally measure elicited responses on evoked emotional responses when listening to music. Specifically, we showcase how we measure the structural music elements as they vary from the beginning to the end within two different compositions, suggesting how and why to analyze and compare standardized, non-personalized behavioral to electroencephalographic data. Reporting our preliminary findings based on this protocol, we focus on the electroencephalographic data collected from n=13 participants in two separate studies (i.e., different equipment and sample background), cross-referencing and cross-validating the biological side of the protocol’s structure. Our findings suggest (a) that all participants – irrespectively of the study – reacted consistently in terms of hemispheric lateralization for each stimulus (i.e., uniform intra-subjective emotional reaction; non-statistically significant differentiation in individual variability) and (b) that diverse patterns of biological representations emerge for each stimulus between the subjects in the two studies (variable inter-subjective emotional reaction; statistically significant differentiation in group variability) pointing towards exogenous to the measurements process factors. We conclude discussing further steps and implications of our protocol approach.

7 144 not been presented so far. In this sense, all aforementioned research protocols may have left 145 aside valuable information that musicology with specific reference to musical structure has 146 to offer in understanding emotionally-evoked responses in the brain and behavior, and 147 whether the musical structure (as a whole -e.g., measurement and time series of a musical 148 piece -or in parts -e.g., a cadenza, appoggiaturas or chord progression) may interact with 149 and alter the neuropsychological system in a standardized way, regardless of the level of the 150 musical experience. This, at least, has already been found to be only the case for the 151 behavioral side of things [38,39].
152 Aims and Hypothesis 153 Considering all the above, the aim of our research project is to measure for the first 154 time in a multimodal way emotional responses to structural elements as they vary from the 155 beginning to the end in a composition (variation measured through continuous response 156 throughout the musical composition) within different compositions, analyzing and 157 comparing standardized behavioral to electroencephalographic data. Our general research 158 hypothesis is that emotion, projecting a particular valence or intensity to a specific (or a set 159 of) structural element(s) of a musical stimulus, provides a joint behavioral and 160 electroencephalographic synchronization imprint to all participants when registered in an 161 ecologically valid testing environment.

162
Following the complexity of the aim and our hypothesis, as well as to be more 163 transparent and concise in the way we explain our findings and Modus Operandi, we 164 decided to present the essential components of our protocol in separate parts -as proof of 165 concept written constructs -making known in detail the process and practical implications 166 that such an endeavor entails. Therefore, in this article, we do not aim to present the final 167 joint findings and analysis of both the behavioral and neuroimaging data, rather provide an 168 in-depth discussion of the conceptual framework of our EEG protocol. A brief presentation 169 of the behavioral part is, however, included. 245 decided to identify the micro-responses in our study. To achieve this, we asked a systematic 246 musicologist ii to perform a music-theoretical analysis. This analysis was cross-referenced 247 and verified by two other professional musicologists, based on aura-based and score-based 248 analyses as well as a combination of two.

249
The music-theoretical analysis identified and converted to time-specific tag points 250 the micro-responses of potential emotional change and produced noted memos in the 251 musical scores presented in Western staff notation. Most importantly, this binary micro-and 252 macro-emotional response approach enabled us to pinpoint and capture inter-subjective as 253 well as intra-subjective differentiations or similarities of neuronal oscillations for each 254 musical composition in the specified timeframes, hence to study emotional responses both 255 at a micro-as well as a macro-neurophysiological time-related-to-emotion frame.

256
The analysis of the Skalkotas composition rendered nine specified timepoints that 257 we refer to as tag points. The Hofmeyr rendered 14. These tag points for each composition 258 are shown in Table 1, presenting the corresponding time duration in minutes and seconds 259 since the beginning of each musical piece. Based on these tag points, we identified the 260 independent and dependent variables of the experiment. The independent variables were 261 our participants and the two musical compositions they listened to, whereas the dependent 262 variables were the hemispheric lateralization of the alpha band activation, as well as the 263 lateralized brain activation effect for and between each tag point.   In order to process and analyze the EEG raw data for both studies, we used the 478 MatLab (Mathworks co.) EEG Lab version 11 software platform. The data were first 479 montaged and segregated into epochs, following first an 'as-per-condition' block 480 segregation and then an 'as-per-tag-points' segregation. A graphical illustration has been 481 already presented in our protocol design earlier on (Fig 1). The following EEG epochs were

487
In more detail, the following configurations/steps were applied in creating and 488 processing these epochs:  Fig 3). These data envelopes were fully extracted from the whole EEG data set for a 507 separate tomographical and statistical analysis as described below. In order to study the hemispheric lateralization dependent variable, we focused on 513 the created epochs (step 7 mentioned above). On these epochs, we followed three steps of 514 analysis for Study 1 (steps 8, 9, and 10) while only one step (step 10) for Study 2. The reason 515 for not following the same detailed path for both studies was that statistical insignificance 516 was found on the mean difference between the two sets of measurements (whole set of 517 electrodes vs. four electrodes; steps 8 and 10) in Study 1, deciding thereafter to exclude 518 steps 8 and 9 from Study 2 (presentation of relevant calculations in Results below).

519
Referring to Study 1, we controlled for the hemispheric lateralization after 520 measuring and analyzing the whole scalp alpha-band activation as projected on all available 521 electrodes and 100% of the EEG data collected for each composition and subject (Step 8). For the aforementioned first case, we followed the approach of the repeated 552 measures ANOVA in order to see how much of the variability shown is a result of the 553 experimental manipulation (tag points in musical stimuli), relative to other random factors 554 (residual) existing in our data collection process (for example, exogenous sound stimulation, 555 prior emotional dynamics etc.) while in the second case, we followed the approach of the 556 independent ANOVA in order to establish if variance differences between participants are 557 isolated or not. In this latter case, the resulting F-test for structuring a sufficient baseline for 558 the overall emotional, biological response for each musical composition can be more potent 559 at a combined sample level.

561
In order to categorize our sample in advance on possible differentiations and 562 variances due to brain structure qualities, we ran an audiometric threshold test using the     Study 2. We investigated the respective alpha band asymmetric activation between 644 hemispheres for each subject and musical composition separately. We first computed the 645 alpha power for the F3, F4, P7 and P8 electrodes as corresponding to the 14 Emotiv+ EPOC 646 unit electrodes system, continuing with a computation of the activation asymmetry ratios Subjects S1 S3 S4 S5 S8 S9 S10 S11 S12 S13 Composers Tags  Seconds  EEG  EEG  EEG  EEG  EEG  EEG  EEG  EEG  EEG The above results present the first evidence of our study for the emotionally induced 711 hemispheric lateralized activity as evoked through a continuous listening process of two