无标度脑波音乐研究

发布时间：2018-06-25 13:54

本文选题：脑电 + 音乐　；参考：《电子科技大学》2013年博士论文

【摘要】：脑电信号（Electroencephalogram，简称EEG）是人脑神经元活动的综合表现，包含丰富的神经信息。音乐是人脑智力活动的产物，对人的身心有着巨大的影响。大脑和音乐之间的关系，一直是神经科学、心理学等领域研究的热点问题。研究如何将脑电信号转换为音乐，并对得到的音乐进行分析和研究，对深入探讨大脑与音乐的关系有重要意义。本文从挖掘脑电信号与音乐的数理共性入手，提出了系列的基于脑电与音乐共同遵循的无标度性质的脑波音乐转换方法，主要工作如下： 1.根据脑电的振幅和音乐中音高的分布都满足的无标度特征，提出了基于单道EEG数据的无标度脑波音乐方法，将EEG信号按照波形特征标记为“事件”，将一个事件映射为一个音符，其中事件的时间长度映射为音长，波形的振幅按照无标度关系映射为音高，平均功率映射为音强。实际的EEG信号被用于音乐生成，结果发现，不同状态下的脑波音乐可以被较好地区分，与另一种基于唤醒度水平的脑波音乐生成方法相比，这种无标度的方法忠实反映了波形的细节信息，并保留了信号原有的无标度性质。同时开发了相应的实时脑波音乐系统，可以用于实时监控或者反馈。 2.在单道无标度脑波音乐方法基础上，提出了一种将两道对称电极的EEG信号转换为两声部合奏脑波音乐的方法。首先将取自左右半球对称电极的信号分别转换为MIDI音乐序列，然后根据中国五声调式和西方七声调式的概念对其进行处理，结果发现音乐的音高分布都符合无标度性，且两种调式在大脑不同状态下得到的标度指数有显著的差异，其中五声调式的音乐声部之间的协和性更高，其和声音程的分布更符合无标度性。另外，经过五声调式的处理，大脑两种不同睡眠状态音乐的差异变得更为明显。 3.通过对大量音乐作品的分析，发现音程的协和性振荡具有无标度性，可以一定程度上反映作曲家的创作偏好，据此建立的音乐家网络可以反映他们之间在和声应用上的传承和相互影响。在此基础上，提出了无标度振荡的脑波音乐方法，首先将两个电极的信号分别转换为MIDI音乐序列，然后利用同样满足无标度振荡的EEG相位同步指数，对该合奏音乐的音程协和性和音强进行调整。当两个信号同步指数高时，音乐协和性就高，音强较大；当同步指数低时，音乐趋向于不协和，音强较小。对安静状态下睁眼和闭眼的脑波合奏音乐进行分析，发现调整后音乐的标度指数更加接近原始脑电信号。这种方法可以对大脑各脑区的协作性的动态变化进行艺术性地表达。 4.分析和讨论了音乐中节奏的功率谱，选择了各种不同风格的中国乐曲，包括古曲、儿歌、流行乐和地方戏曲，发现其节奏功率谱均符合标度分布。在此基础上，提出了多声部脑波音乐方法。在将每个电极的信号分别转换为MIDI音乐序列后，模拟作曲家的创作，设计了艺术滤波器，保留原始音乐中符合调式和节拍要求的音符，最终形成多声部的合奏音乐。结果表明，合奏脑波音乐比单道音乐更具节奏感，其音乐性与丰富性也明显优于单道音乐。此外，合奏脑波音乐还具有更接近原始数据的标度指数，，且更加易于区分不同的状态。 5.将无标度脑波音乐用于齿科正畸疼痛控制。实验中，被试在放置弓丝以后，被分为三个组，施以不同的干预措施。脑波音乐组的被试要求每天聆听本人的脑波音乐，该音乐采用的是在放置弓丝前采集的安静闭眼状态的EEG数据；空白对照组的被试不接受任何干预；认知行为疗法（CBT）组的被试要求每天聆听相应的录音指导。实验结果表明，脑波音乐对于正畸疼痛有较好的控制作用，与空白组和CBT组相比，脑波音乐组在疼痛量表得分上比其他两组有明显的降低，其EEG相干网络的连接密度明显大于空白组，比CBT组也略大，网络参数更接近小世界属性。这说明脑波音乐在疼痛控制上有较为明显的效果，甚至略优于当前热门的行为疗法。
[Abstract]:Electroencephalogram (EEG) is a comprehensive manifestation of the activity of the human brain neuron, which contains rich nerve information. Music is the product of the mental activity of the human brain. It has a great influence on human body and mind. The relationship between the brain and music has always been a hot issue in the fields of neuroscience and psychology. The conversion of electrical signals into music and the analysis and research of the music obtained are of great significance to further explore the relationship between the brain and music.
This paper, starting with the discovery of the mathematical generality of EEG and music, presents a series of scale-free brain wave music conversion methods based on the common compliance of EEG and music. The main work is as follows:
1. according to the scale-free characteristics of the amplitude of electroencephalogram and the distribution of the high sound in the music, a method of scale-free brain wave music based on single channel EEG data is proposed. The EEG signal is marked as "event" according to the waveform characteristics, and an event is mapped to a note, in which the length of the event is mapped to the length, and the amplitude of the waveform is in accordance with no standard. The degree relationship is mapped to pitch, and the average power is mapped to the sound intensity. The actual EEG signal is used for music generation. The results show that brain wave music in different states can be better divided. Compared with another method based on the wakefulness level, this method faithfully reflects the details of the waveform and preserves the information of the waveform. The original scale-free property of the signal is developed, and the corresponding real-time brain wave music system is developed, which can be used for real-time monitoring or feedback.
2. on the basis of single channel non scale-free brain wave music method, a method of converting the EEG signal of two symmetrical electrodes into two voice ensemble brain wave music is proposed. First, the signals from the left and right hemispherical symmetrical electrodes are converted into MIDI music sequences, and then they are carried out according to the concept of Chinese five tone and Western seven tones. The results show that the pitch distribution of the music is in conformity with the scale-free degree, and the scale index of the two modes in different states of the brain is significantly different. Among them, the concordance of the five sound modes is higher, and the distribution of the sound course is more consistent with the standard degree. In addition, the brain is treated with five tones and two different kinds of sleep in the brain. The difference in the sleeping state of music became more obvious.
3. through the analysis of a large number of music works, it is found that the concordance oscillation of the range has no scale, which can reflect the composer's creation preference to a certain extent. On this basis, the network of musicians can reflect the inheritance and mutual influence of the harmonic application between them. On this basis, a method of brainwave music with no scale oscillation is proposed. First, the signals of the two electrodes are converted into MIDI music sequences, and then the synchro and intensity of the ensemble music are adjusted by using the same EEG phase synchronization index that satisfies the scale-free oscillation. When the two signal synchronization index is high, the music concordance is high and the sound intensity is large; when the synchronization index is low, the music tends to be uncoordinated. An analysis of brain wave ensemble music of open and closed eyes in quiet states shows that the scale index of the adjusted music is closer to the original EEG. This method can express the cooperative dynamic changes in the brain regions of the brain.
4. analyze and discuss the power spectrum of rhythm in music, choose different styles of Chinese music, including ancient music, nursery rhyme, pop music and local opera, and find that the power spectrum of the rhythm is in conformity with the scale distribution. On this basis, a multi part brain wave music method is proposed. After converting the signals of each electrode to the MIDI music sequence, The composition of the simulated composer, designed the art filter, retained the notes in the original music that conforms to the requirements of the tone and beat, and finally formed the ensemble music of the multiple voices. The result shows that the ensemble brain wave music is more rhythmic than single music, and its musicality and richness are obviously better than single track music. In addition, the ensemble brain wave music is also more connected. The scaling index near the original data is more easily distinguished from different states.
5. the scale-free brain wave music was used to control the orthodontic pain control. In the experiment, the subjects were divided into three groups after placing the bow wire and were divided into different intervention measures. The subjects of the brain wave group were asked to listen to my brain wave music every day. The music used the EEG data of quiet closed eye state before placing the bow wire; blank control The group's subjects did not accept any intervention; the subjects of the cognitive behavioral therapy (CBT) group were asked to listen to the corresponding recording instructions every day. The results showed that brain wave music had a better control effect on orthodontic pain. Compared with the blank group and the CBT group, the brain wave group was significantly lower than the other two groups in the score of the pain scale, and its EEG coherence. The connection density of the network is obviously larger than that of the blank group, which is slightly larger than the CBT group, and the network parameters are closer to the small world properties. This shows that the brain wave music has a more obvious effect on the pain control, and is even slightly better than the current popular behavioral therapy.
【学位授予单位】：电子科技大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：R318

【引证文献】