Advances in Cognitive Sciences

fa توسعه یک ماشین شناختی جهت ارزیابی عاطفی دستگاه‌های موسیقی ایرانی با استفاده از مدل یادگیری هیجانی مغز Development of a cognitive machine for the evaluation of emotions in the Iranian musical Dastgahs using brain emotional learning مدل سازی شناختی، پردازش سیگنال و تصویربرداری مغز پژوهشي اصیل Research <div style="text-align: justify;">مقدمه: ردیف موسیقی سنتی ایران شامل هفت دستگاه با نام‌های شور، ماهور، سه‌گاه، چهارگاه، راست‌پنجگاه، همایون، نوا است که هر کدام عواطف خاصی را به شنونده انتقال می‌دهد. بسیاری از صاحب‌نظران عرصه موسیقی، ویژگی‌های عاطفی ادراک شده از هر دستگاه موسیقی سنتی ایرانی را بر اساس تجربه به صورت کیفی تشریح کرده اند. ولی نه تنها آنان، بلکه افراد عادی هم در خصوص نوع و اندازه عواطف حاصل از یک دستگاه موسیقی توافق ندارند. زیرا پیش‌زمینه عاطفی افراد متفاوت است. هدف از این مطالعه معرفی یک مدل شناختی است که قادر باشد ویژگی‌های عاطفی هر دستگاه موسیقی را برای شنونده مشخص به صورت کمی پیش‌بینی نماید. روش کار: در این مطالعه یک مدل بهینه یادگیری عاطفی مغز معرفی شد که با مدل روان‌شناختی Thayer ادغام گشت تا بر اساس ویژگی‌های خاص موسیقی دستگاهی ایرانی، الگوی عاطفی دستگاه‌ها را کشف نماید. ما از ردیف میرزاعبدلله به روایت داریوش طلایی استفاده کردیم و از 52 داوطلب خواستیم تا ارزش‌ 12 احساس مدل عاطفی Thayer را برای هر قطعه تعیین کنند. یافته‌ها: برای بررسی عملکرد سیستم، میانگین مربعات خطای سیستم برای داوطلبان بر اساس دستگاه به دست آمد. همچنین با جمع‌بندی نتایج داوطلبان، یک ارزیابی کلی از ویژگی‌های عاطفی دستگاه‌های موسیقی ایرانی ارائه دادیم. نتیجه‌گیری: بررسی سیستم پیشنهادی ما را به این نتیجه می‌رساند که نوعی عدم قطعیت ناشی از پیش‌زمینه ادراکی افراد در عملکرد رفتاری و فیزیولوژیکی احساسات در مغز نهفته است.  </div> Introduction Iranian traditional music includes Dastgahs and Avazes, which can evoke different emotions. Many sources have divided Iranian musical Dastgahs into seven types (Shur, Mahur, Segah, Chahargah, Raste-Panjgah, Homayun, Nava) (&lrm;&lrm;12, &lrm;13). In recent years, several studies have been used for mood recognition or classification of music pieces based on Dastgah (&lrm;14-&lrm;16). Nevertheless, considering this collection wide range of emotions, diversity, and richness, it is necessary to examine the quantitative value of the different emotions evoked by the Dastgah of Iranian music. Famous musicians have qualitatively described the emotional features evoked by each Dastgah. For example, the sadness of the Segah (&lrm;17) and the motivation and pleasure of Mahur (&lrm;18, &lrm;19) have been emphasized. But there are two problems. Initially, a qualitative description is provided, with a limited number of studies focusing on the quantification of emotions evoked by Iranian musical Dastgahs. Secondly, not only musical experts but also ordinary people do not agree with the evoked emotions by each Dastgah. People's emotional backgrounds vary, influencing their comments on a particular Dastgah. Their musical knowledge also plays a role in shaping their opinions. Therefore, musical Dastgah can evoke different emotions in different individuals. The main goal of the current study is to predict the quantitative values of all the emotional features of Iranian musical Dastgahs, considering the emotional background of a specific person and the musical features of an Iranian music piece. Methods The design of the proposed system was based on the parallel processing of emotions in the brain (&lrm;28). The Brain Emotional Learning (BEL) model expresses emotional learning based on the interaction of four neural regions of the brain, including the orbitofrontal, amygdala, thalamus, and cortex (&lrm;25). Thayer's emotional model describes 12 emotions using two dimensions called arousal and valence (&lrm;29). Based on Thayer's model, the emotional studied components are in order: pleased, happy, excited, annoyed, angry, nervous, sad, bored, sleepy, calm, peaceful, and relaxed. According to the proposed structure, pieces of music enter the thalamus, and the thalamus removes the noise, converts the MIDI file to a matrix, and extracts the features based on the structure of Iranian traditional music. The output of the thalamus was sent to all the emotional parts for processing. Based on Thayer's model, the proposed structure included 12 emotional parts, each responsible for evaluating one of the emotions from the input matrix. The most impressive feature was the type of musical Dastgah (including one of the seven Dastgahs) directly imported from the thalamus to the amygdala. Dimensions of the Thayer model helped us to adjust the reinforcement signals. Arousal was used as a signal to the amygdala, and valence as a reinforcement signal to the orbitofrontal. This study used “RADIF of Mirza Abdollah,” published with the composition and playing of Setar (an Iranian musical instrument) by Dariush Talaei (&lrm;31). The researchers asked 52 participants to evaluate each piece’s 12 emotions of Thayer's emotional model. The simulation environment was MATLAB. The cross-validation method was used to evaluate each person's feelings. In this way, the samples were divided into seven equal parts with the normal distribution of parts based on the type of Dastgah, so five parts were used as training samples, and two parts were used as test samples. Twenty-one replications were performed to cover all possible scenarios. Results The proposed system was evaluated in two stages. In the first stage, the system’s overall performance compared to Anfis was reported so that the system’s output was confirmed with an acceptable error. Furthermore, this study used the coefficient of determination or R-Squared as another standard to evaluate the performance of the proposed system. In the second stage, the average emotions of all fifty-two participants were summarized and normalized based on the musical Dastgahs shown in Figure 1. Considering no similar study to compare the system's performance, referring to the qualitative description of the Khaleghi can confirm the results (&lrm;&lrm;17). Fig. 1. Average emotions of all participants based on Dastgah Overall, listeners were more successful in identifying arousal levels in the Dastgahs, but they struggled to discern the valence pattern because arousal is more comprehensible to humans than capacity. The ultimate outcome demonstrated that Dastgahs with higher arousal power, such as Shur and Mahur, had a lower average system error. Conversely, Dastgahs with lower valence power, like Nava and RastePangah, experienced a higher error rate. Conclusion                                This study was conducted to evaluate individual emotions of Iranian musical Dastgahs. This research tried to propose a model with special attention to the features of Iranian traditional music and the emotional background of the individual. The proposed model included 12 parts, each responsible for processing one emotion of Thayer's emotional model. All 12 parts worked in parallel with each other. Each part included four areas: orbitofrontal, amygdala, thalamus, and cortex. The most impressive feature was the type of Dastgah that entered the amygdala directly from the thalamus. Thayer's model has two dimensions: valence and arousal. The proposed model used valence to regulate the amygdala reinforcement signal and arousal to regulate the orbitofrontal reinforcement signal. The analysis of the results leads us to the existence of uncertainty caused by people's emotional background lies in the behavioral and physiological function of emotions in the brain. Ethical Considerations   Compliance with ethical guidelines  The present study was conducted following ethical principles, including the consent of all participants, respect for the confidentiality of participants' information, and freedom to leave the research process.   Authors' contributions All four authors prepared the study; the first author designed the model and the computational framework. The second author analyzed the data. The third and fourth authors did sample preparation. All of the authors contributed to the interpretation of the results.           Funding  No financial support has been received from any organization for this research. Acknowledgments                 The authors would like to thank all 52 participants in this study, as well as the respected professors who guided this work. Conflicts of interest All authors certify that they have no affiliations with or involvement in any organization or entity with any interest or non-financial interest in the subject matter or materials discussed in this manuscript.   مدل‌سازی شناختی, ارزیابی عاطفی, دستگاه موسیقی ایرانی, یادگیری عاطفی مغز, ردیف میرزاعبدالله Cognitive modeling, Evaluation of emotions, Iranian musical Dastgah, Brain emotional learning, RADIF of Mirza Abdollah 108 122 http://icssjournal.ir/browse.php?a_code=A-10-1593-1&slc_lang=fa&sid=1 Maryam Jandaghian مریم جندقیان 100319475328460015362 100319475328460015362 No PhD Student of Artificial Intelligence, Department of Computer Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran دانشجوی دکتری هوش مصنوعی، دانشکده فنی و مهندسی، دانشگاه آزاد اسلامی واحد علوم و تحقیقات، تهران، ایران Saeed Setayeshi سعید ستایشی Setayesh@aut.ac.ir 100319475328460015363 100319475328460015363 Yes Associate Professor, Department of Energy Engineering and Physics, Amirkabir University of Technology, Tehran, Iran دانشیار گروه مهندسی پرتو پزشکی، دانشگاه صنعتی امیرکبیر، تهران، ایران Farbod Razzazi فربد رزازی 100319475328460015364 100319475328460015364 No Associate Professor, Department of Electrical Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran دانشیار دانشکده مکانیک، برق و کامپیوتر، دانشگاه آزاد اسلامی واحد علوم و تحقیقات، تهران، ایران Arash Sharifi آرش شریفی 100319475328460015365 100319475328460015365 No Assistant Professor, Department of Computer Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran استادیار دانشکده مکانیک، برق و کامپیوتر، دانشگاه آزاد اسلامی واحد علوم و تحقیقات، تهران، ایران