Tagore and neuroscience: A non-linear multifractal study to encapsulate the evolution of Tagore songs over a century

https://doi.org/10.1016/j.entcom.2020.100367Get rights and content

Highlights

  • 4 Tagore songs sung by artistes for over five generations have been taken.

  • Multifractal analysis of the songs carried out for different generation of singers.

  • The singing style of a singer/generation has been quantified by multifractal width.

  • The neural response of the evolution is assessed with the help of EEG.

  • EEG complexity provides information about valence lateralization in different lobes.

Abstract

The verses of Rabindranath Tagore have been sung by various artistes over generations spanning over almost 100 years. There are few songs which were popular in the early years and have been able to retain their popularity over the years while some others have faded away in the course of time. In this study we tried to find cues in the singing style of these songs, sung by different singers spanning over almost five generations, which have kept them alive for all these years. For this, we took 3 min clips of four Tagore songs which are being sung by atleast five generations of artistes over 100 years and analyzed the acoustic signals with the help of latest nonlinear technique Multifractal Detrended Fluctuation Analysis (MFDFA). Next EEG data was collected from 5 persons who listened to 30 sec clips of two Tagore songs sung over five generations of artistes in chronological order. The EEG response from the participants were analyzed with the help of the same MFDFA technique and the multifractal spectral width was considered as the parameter which can help in the identification of cognitive evolution of the Tagore songs. The multifractal spectral width is a manifestation of the inherent complexity of the signal and in future, may prove to be an important parameter to identify the singing style of a particular generation of singers and how this style varies over different generations. The EEG responses from the participants reflect how the perception and cognition of the same Tagore songs evolve over generations. The results and implications are discussed in detail.

Introduction

The great visionary from Bengal, Rabindranath Tagore once said “Whatever fate may be in store in the judgment of the future for my poems, my stories and my plays, I know for certain that the Bengali race must accept my songs, they must all sing my songs in every Bengali home, in the fields and by the riversI feel as if music wells up from within some unconscious depth of my mind, that is why it has certain completeness” (Tagore, Bangla 1407). More than 75 years after his demise, we know for sure that his songs have sustained the various changes that have come in our society and been modified accordingly to keep itself relevant. Tagore (1861–1941) was a Bengali poet, philosopher, artist, playwright, composer and novelist. India's first Nobel laureate, Tagore won the 1913 Nobel Prize in Literature for Gitanjali or Song Offerings. He composed the text of both India's and Bangladesh's respective national anthems. Tagore travelled widely and was friends with many notable 20th century figures such as William Butler Yeats, H.G. Wells, Ezra Pound, and Albert Einstein. Accordingly, a number of his songs had been influenced by amalgamation from a number of different cultures. While caring for both the traditions, classical and folk, he respected the inviolable sanctity of neither and freely took from each what suited his purpose. He was not even averse to borrowing from western melodies, although he did very little of that and made his own whatever he took from other sources. Tagore's music cannot be separated from his literature, most of which went on to become lyrics for his songs. Primarily influenced by the thumri style of Hindustani classical music, Tagore’s songs span the entire spectra of human emotion, ranging from his early dirge-like Brahmo devotional hymns to quasi-erotic compositions [11]. They emulated the tonal color of classical ragas to varying extents, while at times his songs mimicked a given raga's melody and rhythm faithfully, he also blended elements of different ragas to create innovative works [7]. For Bengalis, their appeal—arousing from the combination of emotive strength and beauty described as surpassing even Tagore's poetry—was such that the Modern Review [6] observed that “there is in Bengal no cultured home where Rabindranath's songs are not sung or at least attempted to be sungEven illiterate villagers sing his songs.” Such is the impact of Tagore songs that a recent study by The Telegraph reveals Rabindra Sangeet collections remain on the best selling lists of various music stores even now, one in every four cassettes sold is a Rabindrasangeet album while on CD, the ratio is one in three [18]. What is the secret formula which makes Tagore songs, which were composed nearly 100 years back, survive and flourish in this age of computer generated music? There have been very few studies [39], [9], [32], [45], [46] which try to scientifically analyze the evolution of Tagore songs over the years. This is somewhat surprising as Tagore himself believed in the scientific analysis of the expression of various melodies and its impact on human mind [28].

In the recent past, there have been a number of studies which tried to look into the connection between Tagore songs and neuroscience [27], [28]. With the advent of technology and advances in human-computer interaction, explaining "How the brain dictates the mind?" has become one of the major researhes of neural studies. A number of works in Western classical music tries to capture the evolution of style using different modes of music information retrieval [51], [16], [35], but the brain response of this kind of evolution is seldom studied in the Indian perspective. How does the activity of discrete neurons achieve awareness for us, for example, experiencing gladness in the heart? Discoveries and answers coming from neuroscience quarters did not miss the attention of philosophers and theologians, and at this juncture the scientific study of Tagore's works and its connection with neuroscience becomes very relevant. In [27], the author briefs on how Gitanjali anticipates the view of a person who is part of Tagore’s universe of discourses and that some of his later prose writings [47] sketch. It runs counter to naturalistic assumptions, thus challenging claims based on discoveries of electrical activities and the biochemistry of synaptic conduction that hold out the possibility to explain the full story of being a human person, or, more specifically, our experience of joy, mystery, and the transcendent in terms of the natural. The author also briefs on how the choice of words such as “bound” may give us the idea of personness of Tagore. Lines with such occurrences in general convey the idea that becoming a person is bound-up with and to a transcendent power that limits or defines our existence and thus make us as individuals more than simply materiality. Each occurrence underscores slightly differently that very idea of personness. The classical Indian view on artistic creativity, shared and elaborated on by Tagore, is almost silent on the kind of unconscious processes that form the forte of modern psychological and neuroscience research. The Indian view states that creative imagination has access to a transcendent, spiritual unconscious that lies in a deeper layer of the mind than the strata in which the cognitive, dynamic and other unconscious processes are located. Tagore attributed his creativity to this spiritual unconscious which he calls 'One Within Me' and whose creations, such as pictures, poems, music are pastime through which it finds joy. The recent upsurge in scientific research on Tagore is particularly centered amongst researchers for whom understanding human cognition is imperative to creating artificial neural network based system of cognition. For Tagore, boundary is an important term in his universe of discourse. This was acknowledged in a special issue of the journal Annals of Neuroscience [47], published on the occasion of 150th birth anniversary of Tagore which was based on Gitanjali, as the Neuroscience of Music. In all these years, when the songs and writings of Tagore have been a source of immense pleasure, least knowledge have been acquired about the cognitive appraisal of the structure of these songs which have stood the test of time. In this study for the first time we look to unveil the complex structure of the evolving Tagore songs as well as look for neural signatures of this evolution spanning a century.

The fractal and multifractal aspects of different genres of music were analyzed by Bigerelle and Iost [4], where it was proposed that the use of fractal dimension measurements could benefit the discrimination of musical genres. Su and Wu [44] applied Hurst exponent and Fourier analysis in sequences of musical notes and noted that music shares similar fractal properties with the fractional Brownian motion. But, music signals may exhibit self similarity in different scales which may not be described by single scaling exponent as is found in Detrended Fluctuation Analysis (DFA) [37] technique. That is, the clustering pattern is not uniform over the whole system. Such a system is better characterized as ‘multifractal’. Recent research with complex systems showed that naturally evolving geometries and phenomena cannot be characterized by a single scaling ratio (as in monofractal system), as different parts of the system are scaled differently. Such a system is better characterized as a multifractal system. A multifractal can be loosely thought of as an interwoven set constructed from sub-sets with different local fractal dimensions. Real world systems are mostly multifractal in nature. Music originated in the sounds of nature, and hence like many other naturally occurring fluctuations, music also has a fractal structure. In a number of recent works, nonstationarity and nonlinearity of the time series of music signals have been used to quantify the complexity/ musicality of the acoustic signal as part of multifractal analysis [40], [34], [20], [43], [22], [53]. As almost every music signal has naturally evolving geometry and non-uniform pattern in its progression, multifractal formalism can prove to be an important tool which delves into the finer intricate details of the music signal.

Since the multifractal technique analyzes the signal in different scales, it is able to decipher much accurately the amount of self similarity present in a signal. The spectrum in multifractal detrended fluctuation analysis (MFDFA) [25] is the measure of complexity or self-similarity present in the signal. This method is very useful in the analysis of various non-stationary time series and it also gives information regarding the multifractal scaling behaviour of non-stationary signals. The acoustic waveform arising out of different Tagore songs are no different and the multifractal technique is expected to decipher information from them at the deepest level of accuracy. We used this technique to characterize the collected music samples (different Tagore songs sung over the century) and classified them on the basis of their complexity values which is determinant of the uniqueness of that particular song. Furthermore, the neruronal signatures of this evolution have been characterized with the help of EEG signal analysis using these songs as input stimuli.

In the next part of the study, we look into the neuro-scientific tools to assess how the cognitive appraisals of Tagore songs evolve over the period of 100 years. The Electroencephalography (EEG) data is a time variation record of the neuronal fluctuations occurring over time in the human brain. It is essentially a neuro-scientific bio-sensor which provides plentiful information about the complex human brain dynamics according to electrical activity in brain tissues (waves i.e. plot of voltage difference over time between electrodes by using the summation of many action potentials sent by neurons in brain) against human emotion elicited by music. The scalp EEG arises from the interactions of a large number of neurons whose interactions are generally nonlinear and thus they can generate fluctuations that are not best described by linear decomposition. In recent past, the DFA has become a very useful technique to determine the fractal scaling properties and long-range correlations in noisy, non-stationary time-series. It has been widely applied to diverse fields such as DNA sequences, heart rate dynamics, neuron spiking, human gait, and economic time-series and also to weather related and earthquake signals [33], [5], [37], [2]. DFA has also been applied to EEG signals to identify music induced emotions in a number of studies [15], [26], [3]. Gao et al. [15] related emotional intensity with the scaling exponent, while a recent study [3] relate the variation of alpha scaling exponent generated from DFA technique with the retention of musical emotions – an evidence of hysteresis in human brain. But DFA has its own limitations. Many geophysical signals as well as bio-signals do not exhibit monofractal scaling behavior, which can be accounted for by a single scaling exponent [19], [23], therefore different scaling exponents are required for different parts of the series [8]. Consequently a multifractal analysis should be applied. The Multifractal Detrended Fluctuation Analysis (MFDFA) technique was first conceived by [25] as a modification of the standard DFA. MFDFA has been applied successfully to study multifractal scaling behavior of various non-stationary time series [38], [24], [52], [50] as well as in detection or prognosis of diseases [12], [13], [17]. The multifractals are fundamentally more complex and inhomogeneous than monofractals [42], [48] and can describe time series featured by very irregular dynamics, with sudden and intense bursts of high-frequency fluctuations [31]. EEG signals are essentially multifractals as they consist of segments with large fluctuations as well as segments with very small fluctuations, hence when applied to the alpha and theta EEG rhythms, the multifractal spectral width will be an indicator of emotional arousal corresponding to a particular clip. In case of music induced emotions, a recent study [29] used the multifractal spectral width as an indicator to assess emotional arousal corresponding to a simple musical stimuli – a tanpura drone. Thus, it is expected that EEG signal acquired with different generations of Tagore songs as stimuli will bear some reflection on the cognitive evolutionary process of the songs. How do the cerebral perception of the Tagore songs evolve or vary over the century is a question of immense importance today. This study may throw some light into the neuro-cognitive significane that Tagore’s works have to offer even in today’s scenario.

Four widely popular songs of the maestro Rabindranath Tagore were taken for our analysis, which have been popular over the years, sung by singers of different generations. The songs are Bohujuger Opar Hote, Maharaj Eki Saje, Krishnakoli Ami Tarei Boli and Amaro Porano Jaha Chay. The rationale behind choosing these songs is that they are among those few compositions which have sustained through the ages and is still being sung by singers of this era, while many others have lost their popularity. The multifractal width (or the complexity values) was computed for each of the phrases and their variation was studied over the years to identify cues for evolution. The study reveals interesting results regarding the style of rendition of each individual performer or a group of performers of the same generation. For e.g. while the singers of the first generation (when there were no formal musical notation of the songs) showed significantly high values of complexity throughout their rendition, singers of the latter generations showed a decrease in the complexity values. Another interesting observation is that, those singers who have been trained in Vishwa Bharati or under the patronage of the maestro himself show a distinct lineage or unique style of rendition clearly distinguishable from the others. Also, with the help of MFDFA technique, we have quantitatively evaluated the brain response corresponding to the five clips in six different electrodes from three different lobes namely, F3/F4 (frontal lobe), T3/T4 (temporal lobe) and O1/O2 (occipital lobe). The frontal electrodes have been reported to be associated with higher order cognitive thinking, while temporal electrodes have been associated with auditory skills while occipital lobes are generally associated with visual cognition. Keeping these rationales in mind, we decided to consider the EEG fluctuations corresponding to these six electrodes while the participants listened to clips of Tagore songs pertaining to five different generations. Thus, with this study, we have tried to quantify and visualize how singers have rendered the same song in their unique style of singing so as to keep the popularity value intact but without losing the flavor of Rabindrasangeet. The change in the singing style of the same song over the years by different artistes is also manifested in this study.

We have used MFDFA technique to measure temporal variation of self-similarity in the audio waveforms of various Tagore songs to reveal the internal dynamics of their so-called musicality. The width of the multifractal spectrum for the waveform of each audio sample is measured and is ascribed to be the parameter with which the melodic evolution of the Tagore songs for over a century has been quantified. The multifractal spectral width 'W' essentially measures the length of the range of fractal exponents in the audio signal; therefore, wider the range, the ‘‘richer’’ the signal in structure. In this way, W evolves to be an unique parameter with which the singing pattern of a particular generation of singers have been quantified. Recent studies [49], [30] report that the stimulus with intermediate level of complexity is more aesthetic than the stimulus with higher or lower level of complexity. By complexity, though, here we mean objective complexity which is computed mathematically by different robust methods. Thus aesthetic appreciation of the singing arrangement of different singers might provide a correlation with the obtained multifractal complexity of the acoustical signal of the song.

Section snippets

Subjects summary

5 (M = 3, F = 2) musically untrained adults, chosen randomly from the pool created for listening test data, voluntarily participated in this study. The average age was 25 years (SD = 4.35 years) and average body weight was 70 kg. Informed consent was obtained from each participant according to the guidelines of the Ethical Committee of Jadavpur University. All experiments were performed at the Sir C.V. Raman Centre for Physics and Music, Jadavpur University, Kolkata. The experiment was

Methodology

The raw EEG signal is generally contaminated by various types of external artifacts such as eye blinks, muscular movement etc. Eye blinks and eye movements are characterized by frequency of less than 4 Hz and high amplitude. Thus, it is essential to identify these artifacts and to remove them from the raw EEG signal to get a noise free EEG data. A novel data driven technique called Empirical Mode Decomposition (EMD) was used to de-noise the EEG signal [29]. The noise-free EEG signals

Results and discussion

In the beginning, we start with the observations from the acoustic analysis and continue with the results obtained from EEG analysis. Every musical composition/element can be considered as a nonlinear complex time series – the multifractal width (W) being a quantitative measure of its complexity. In other words, more W – more local fluctuations in temporal scale and thus this parameter is very useful to characterize and quantify a particular music signal up to a level which is not possible by

Conclusion

This abiding relevance of Rabindrasangeet through the ages is its hallmark as a literary masterpiece. It was in this context that Tagore felt very grateful for the efforts made by Dutch musicologist Arnold Bake after he left Santiniketan, to familiarise European audiences with his music. Rabindranath was conscious that his songs could not find easy recognition in the west. Even within India, in some pockets of the non Bengali speaking world, while Tagore’s songs have a certain appeal for their

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

One of the authors, AB acknowledges the Department of Science and Technology (DST), CSRI, Govt. of India for providing (CSRI-PDF/34/2018) the CSRI Post-Doctoral Fellowship to pursue this research work. Another author, SS acknowledges the JU RUSA 2.0 Post-Doctoral Scholarship Scheme for providing the Post-Doctoral Fellowship to pursue this research (R-11/557/19).

References (52)

  • S. Sanyal et al.

    A study on improvisation in a musical performance using multifractal detrended cross correlation analysis

    Physica A

    (2016)
  • E. Serrano et al.

    Wavelet leaders: a new method to estimate the multifractal singularity spectra

    Physica A

    (2009)
  • H.E. Stanley et al.

    Statistical physics and physiology: monofractal and multifractal approaches

    Physica A

    (1999)
  • Z.Y. Su et al.

    Multifractal analyses of music sequences

    Physica D

    (2006)
  • Z.Y. Su et al.

    Music walk, fractal geometry in music

    Physica A

    (2007)
  • L. Telesca et al.

    Mono-and multi-fractal investigation of scaling properties in temporal patterns of seismic sequences

    Chaos, Solitons Fractals

    (2004)
  • N.K. Vitanov et al.

    Nonlinear time series analysis of vibration data from a friction brake: SSA, PCA, and MFDFA

    Chaos, Solitons Fractals

    (2014)
  • L. Zunino et al.

    A multifractal approach for stock market inefficiency

    Physica A

    (2008)
  • S. Roy et al.

    Chaos based non-linear cognitive study of different stimulus in the cross-modal perspective

    Physica A: Statistical Mechanics and Its Applications

    (2020)
  • Y. Ashkenazy et al.

    Nonlinearity and multifractality of climate change in the past 420,000 years

    Geophys. Res. Lett.

    (2003)
  • Y. Ashkenazy et al.

    Magnitude and sign correlations in heartbeat fluctuations

    Phys. Rev. Lett.

    (2001)
  • M. Chatterjee

    Tagore and Jana Gana Mana

    Counter Currents

    (2003)
  • A. Chaudhuri

    The Vintage Book of Modern Indian Literature

    (2004)
  • Z. Chen et al.

    Effect of nonstationarities on detrended fluctuation analysis

    Phys. Rev. E

    (2002)
  • A.K. Datta, S.S. Solanki, R. Sengupta, S. Chakraborty, K. Mahto, A. Patranabis, Signal Analysis of Hindustani Classical...
  • K. Dutta et al.

    Rabindranath Tagore: The myriad-minded man

    (1995)
  • Cited by (11)

    • On the application of deep learning and multifractal techniques to classify emotions and instruments using Indian Classical Music

      2022, Physica A: Statistical Mechanics and its Applications
      Citation Excerpt :

      Usually, MER becomes a matter of huge challenge especially because of its ambiguous emotional response. Studies involving non-linear techniques have been conducted in the recent past to understand this complex behavior of ICM and its manifestation in the human brain [38–49]. MER, particularly with ICM is still an uncharted territory for researchers especially because of the lack of structured datasets comprising of ICM clips.

    • A novel study on perception–cognition scenario in music using deterministic and non-deterministic approach

      2021, Physica A: Statistical Mechanics and its Applications
      Citation Excerpt :

      These studies have revealed that music signals feature a very complex behavior as at every instant components (both in micro and macro scale) like pitch, timbre, accent, note sequence, melody etc are closely linked to each other [48,49]. A number of studies have shown the non-stationary and nonlinear nature of music signals [50,51] and hence a non-deterministic/chaotic approach is needed in understanding the music signals [39,52,53]. In this context fractal analysis of the music signal seems a suitable tool as it reveals the intrinsic geometry embedded in the signal by measuring the degree of self similarity present in the chaotic non-linear system.

    View all citing articles on Scopus

    This paper has been recommended for acceptance by Matthias Rauterberg. (Where JONG-IL PARK is the name of the Handling Editor).

    View full text