Exploring the Effects of Gratitude Voice Waves on Cellular Behavior: A Pilot Study in Affective Mechanotransduction

Abstract

Emotional communication is a multi-modal phenomenon involving posture, gestures, facial expressions, and the human voice. Affective states systematically modulate the acoustic signals produced during speech production through the laryngeal muscles via the central nervous system, transforming the acoustic signal into a means of affective transmission. Additionally, a substantial body of research in sonobiology has shown that audible acoustic waves (AAW) can affect cellular dynamics. This pilot study explores whether the physical–acoustic changes induced by gratitude states in human speech could influence cell proliferation and Ki67 expression in non-auditory cells (661W cell line). We conduct a series of assays, including affective electroencephalogram (EEG) measurements, an affective text quantification algorithm, and a passive vibro-acoustic treatment (PVT), to control the CO₂ incubator environment acoustically, and a proliferation assay with immunolabeling to quantify cell dynamics. Although a larger sample size is needed, the hypothesis that emotions can act as biophysical agents remains a plausible possibility, and feasible physical and biological pathways are discussed. In summary, studying the impact of gratitude AAW on cell biology represents an unexplored research area with the potential to enhance our understanding of the interaction between human cognition and biology through physics principles.

1. Introduction

Since the earliest recorded history, emotions have been the subject of interest for the scientific community due to their significant impact on humans’ daily lives and well-being [1]. Emotions are an inherent aspect of the human experience and play a fundamental role in communicating with others [2], affecting posture, gestures, and facial and oral expressions [3,4,5,6,7,8]. However, although several studies have attributed a prominent role to the voice in the recognition of emotional experience, oral expressions of emotion have been considerably less extensively investigated than facial expressions [9,10].

The articulation of emotional states through sound is a complex phenomenon. People raise their voices when they are happy or angry, modifying their voice characteristics through the laryngeal muscles [11,12] and encoding emotional information in the resulting speech signal [13,14,15]. Ultimately, emotions modify the characteristics of the phonatory apparatus [13,14,15], affecting different acoustic parameters such as fundamental frequency contour (F0), timbre, amplitude, and temporal aspects [16,17,18,19]. It has been demonstrated that humans can perceive and differentiate the emotions of others by listening to voice signals [20,21,22,23,24].

One emotion that falls within the positive emotion category is gratitude. Consider the affective feeling of having a flat tire on a remote road and a passerby stops to assist you in changing it. Such unexpected acts of kindness can evoke strong feelings of gratitude, emphasizing the importance of social support in times of need. Gratitude is strongly associated with higher well-being, social relationship benefits, and lower stress and depression levels [25,26]. Several studies have revealed that gratitude correlates with areas of the anterior cingulate cortex (ACC) and medial prefrontal cortex (mPFC) [27,28]. Electroencephalogram (EEG) devices have demonstrated their potential for classifying positive emotions, including gratitude, with an accuracy of approximately 80% [29]. Likewise, the use of gratitude letters has been demonstrated to be an effective method for inducing feelings of gratitude in laboratory settings [30,31], and affective dictionaries have been proposed as a systematic method for assessing and quantifying the emotional content of textual data [32,33].

Additionally, several studies have demonstrated the capacity of audible acoustic waves (AAW) to alter cell behavior [34,35]. The well-established field of mechanotransduction examines the physical–biological process by which AAWs reach the extracellular matrix and interact with cell nuclei [36,37,38]. Building on this idea and the concept that emotions act as physical modulators of human speech, we hypothesize that gratitude AAWs could influence cellular dynamics through the human voice. Specifically, this pilot study aims to determine whether gratitude, as a biophysical agent, can affect the proliferation of non-auditory cells.

2. Materials and Methods

2.1. Participants

The participant, a 48-year-old Spanish right-handed male, was a renowned spiritual guide. This study was designed in accordance with the Declaration of Helsinki and was approved by the Institutional Committee for Ethical Review of Projects (CIREP-UA) of the University of Alicante (Alicante, Spain).

2.2. Design

Four different AAWs were employed during experimentation: (i) control, (ii) emotional voice, (iii) non-emotional voice, and (iv) synthesized emotional voice. At the beginning of the session, the participant wrote a gratitude letter in accordance with the Toepfer and Walker protocol [39] in a quiet, private room.

The gratitude letter (see Appendix A) was then read out loud in front of a transductor to generate the emotional voice condition (ii). The non-emotional voice condition (see Appendix A) was a recording firstly of a text fragment from Don Quixote de la Mancha, under identical conditions as in case (ii), by the same subject. Non-emotional stimuli (iii) were recorded before emotional stimuli (ii) because the effects of writing a gratitude letter may be persistent over time [30]. The synthesized emotional voice condition (iv) was synthesized from the gratitude letter written by the participant, and it contained the same words and cadence of that condition (ii). During the readings (ii) and (iii), the participant’s EEG recording was taken. The total time spent in the room (writing gratitude letter + placing EEG electrodes + recording emotional and non-emotional stimuli) was about 45 min. Subsequently, a word analysis algorithm using affective dictionaries was developed to quantify the emotional component of texts.

Ultimately, an electro-acoustic radiation system based on a loudspeaker and passive acoustic treatment was installed in the CO₂ incubator. All AAWs were radiated over non-auditory cells for 72 h in independent experiments at a distance of 27 cm. In the control condition (i), the electroacoustic system remained off. Biological assays were carried out.

2.3. Set-Up

2.3.1. Audio Recording System

The recording system consisted of a condenser microphone (model Rode NT5 S, produced by RØDE, Sydney, Australia), situated at a distance of 15 cm from the source, coupled with a passive acoustic device (model Kaotica Eyeball, Winnipeg, MB, Canada) to minimize the impact of the recording environment. The transducer was connected to a Zoom-H6 recorder and registered in WAV format (16-bit, 44,100 Hz).

2.3.2. EEG Data Recordings

EEG signals were captured using 15 passive electrodes (FP1, FPz, F7, F3, Fz, F4, F8, C3, Cz, C4, P3, Pz, P4, O1, O2) arranged in accordance with the international 10–20 system [40]. The recordings were taken using the OpenBCI system, with a sampling rate of 125 Hz, and touch-proof electrode adapters (all produced by OpenBCI, Brooklyn, NY, USA). Electrode impedance was maintained below 50 kOhms, according to the impedance mode coded in the software of the OpenBCI GUI (version 5.2.2). The default settings of the OpenBCI electrode cap were used for the online reference and ground electrodes.

2.3.3. Passive Vibro-Acoustic Treatment

The reason for using a passive vibro-acoustic treatment (PVT) was twofold. On the one hand, since a diffuse field can be assumed in an enclosure, the reverberant field is expected to dominate the acoustic behavior in a CO₂ incubator. On the other hand, passive sound-absorbing materials and insulating materials reduce background noise generated by electro-mechanical systems and actuators improving signal-to-noise ratio (SNR) while homogenizing the acoustic environment at the same time. For this purpose, EliAcoustic Pyramidal sound-absorbing material (595.0 × 595.0 mm with a thickness of 50 mm and a density of 29 kg/m³) was placed on the walls of the incubator (FormaTM Steri-CycleTM CO₂ incubator, model 371, Thermo Electron Corporation, Waltham, MA, USA), and COPOPREN insulation (470 × 490 mm, 80 mm thick and 80 kg/m³ density) was installed in the incubator tray to reduce structural vibrations (Figure 1a).

2.3.4. Electro-Acoustic Radiation System

The acoustic stimulus was radiated onto the non-auditory cells using a Bose SoundLink Color II loudspeaker suspended on an acoustic hanger in a CO₂ incubator at an equivalent sound pressure level (L_p) of 80 dB (measured at 1 min intervals and calibrated with an NA-28 sound level meter, serial number 00762289, Rion Co., Ltd., Tokyo, Japan) for 72 h (Figure 1b). The transducer was connected to a digital audio player (AGPTEK 32 GB, manufactured by AGPTEK Shenzhen, China) with a properly shielded 3.5 mm mini-jack cable and a HUB supplying power.

2.3.5. Noise and Vibration Measurements

Experimental measurements were employed to characterize the set-up, which was composed of ARTA software (version 1.9.1), a TASCAM 144 mk2 acquisition card, and a Bose SoundLink Color II loudspeaker suspended on an acoustic hanger. The measurements were taken with a Brüel & Kjaer microphone transducer (4188-A-21, B&K, Copenhagen, Denmark) connected to a 2-channel Brüel & Kjaer CCLD signal conditioner (1704-A-002, B&K, Copenhagen, Denmark) with Z-weighting for a 1 min duration (intervals of 0.1 s). To measure the sound pressure level (L_p), a Brüel & Kjaer microphone (Type 4188-A-21, B&K, Copenhagen, Denmark) was used.

2.4. Data Analyses

2.4.1. Voice Synthesis Method

The synthesized emotional signal (iv) was generated from the emotional voice (ii) using VoiceOver, a reading technology based on voice synthesis integrated into the Mac OS X operating system. Starting with the gratitude letter written by the participant, the synthetized audio signal was adjusted by following the wave-log in an audio sequencer (Steinberg Cubase LE 11.1 software) to ensure maximum synchrony and balancing of amplitudes.

2.4.2. Audio Filtering and Processing

All audio signal was processed with a passband filter (150 Hz and 10 KHz cutoff frequencies), normalized to −3 dB, and digitally limited to −1 dB. A noise gate was employed to minimize the unwanted sound artifacts.

2.4.3. Audio Analysis Magnitude-Squared Coherence

Magnitude-squared coherence (MSC) was used to measure the similarities between the spectral content of (ii) the emotional signal and (iv) the synthesized emotional signal [41]. The MSC of the two signals, x(t) and y(t), was defined as [42]:

$$C_{xy}\left(f\right)={\displaystyle \frac{{\left|P_{xy}\left(f\right)\right|}^2}{P_{xx}\left(f\right)·P_{yy}\left(f\right)}}$$

(1)

where $P_{xy}\left(f\right)$ represents the cross power spectral density and $P_{xx}\left(f\right)$ and $P_{yy}\left(f\right)$ denote the power spectral density. The MSC was calculated for frequencies between 20 and 5000 Hz (mscohere function in MATLAB).

2.4.4. Audio Spectrogram

In order to analyze the temporal variation of (ii) the emotional signal and (iv) the synthesized emotional signal, a spectrogram was calculated (spectrogram function in MATLAB).

2.4.5. Affective Text Algorithm

Affective dictionaries can estimate the writer’s feelings, providing a computational approach to investigating the emotional connotations of written texts. Two databases have been used to check whether writers are displaying gratitude feelings: the lexicon of the National Research Council of Canada built on parameters of Valence, Arousal, and Dominance (NRC-VAD) and the NRC Word-Emotion Association Lexicon (EMOLEX) [33,43]. The NRC-VAD Lexicon is a dictionary based on a Russell tridimensional model that includes 19,917 English words classified by valence, arousal, and dominance [44]. The EMOLEX, on the other hand, is a dictionary that includes 14,154 English words classified by feeling (positivity and negativity) and emotions (anger, anticipation, disgust, fear, joy, sadness, surprise and trust). To test whether there was a gratitude feeling in the participant’s voice, an affective text analysis was conducted to quantify the valence and arousal levels associated with both (i) emotional and (ii) non-emotional stimuli.

2.4.6. EEG Pre-Processing and Filtering

An EEG recording was conducted in order to compare the participant’s ACC activity during the recording of (ii) emotional (gratitude) and (iii) non-emotional stimuli (resting-state). We selected a fixed data segment from all signals, using the first 2.244 min (maximum amount of data in one of the conditions) for analysis. Data were linearly demeaned. To prevent edge artifacts during filtering [45], we reflected the first and last 10 s of each signal, which were trimmed after filtering. The EEG data were then zero-phase filtered (filtfilt function in MATLAB), with a 4th-order Butterworth high-pass filter set at 0.5 Hz.

2.4.7. Estimating the Power Spectral Density (PSD) for Each Frequency Band

We estimated the power spectrum density (PSD) for each data segment within the broadband range of 1–45 Hz at the Fz electrode using a modified version of the Welch method [46] (window = 1 s; 0% overlap; resolution = 1 Hz). To minimize autocorrelation between contiguous windows, our modification only considered a single 1 s window every three 1 s windows (i.e., 1 s every 3 s, as computed in [47]). Each condition had a total of 45 windows of 1 s. Power values were log-transformed and converted to the unit 10*log10[μV²]. For each frequency, the spectral power was detrended for all 1 s spectral estimates throughout the full recording before comparing conditions. This approach minimized the impact of spectral power drifts in the signal, preventing artificial differences between conditions. The resulting spectral power values were computed using Simpson’s rule (bandpower function in MATLAB) in the theta (4–7.9 Hz), alpha (8–11.9 Hz), beta (12–29.9 Hz), and gamma (30–45 Hz) frequency bands. We excluded the delta band (1–3.9 Hz) from the analysis due to its susceptibility to blinking and movement artifacts [48].

2.4.8. EEG Statistical Analysis

Before computing any statistical analyses, we checked for potential outliers in our data (power of in each of the 1 s windows) by running the Grubbs test with an α-level of 0.05 (isoutlier in MATLAB). If any significant outliers were found, they were removed. We checked for data normality using the Kolmogorov–Smirnov test (kstest in Matlab). For statistical assessment of the power values in each frequency band, we applied a two-tailed paired t-test (t-test in MATLAB) and corrected the p-values for multiple comparisons by applying the false discovery rate (FDR) of Benjamini and Hochberg [49] (fdr_BH function in the Multiple Testing Toolbox [50]). For the statistical assessment of the power values across the frequency range, we performed sample-by-sample paired t-tests between the spectral power at the Fz electrode between resting-state and gratitude conditions. We used a cluster-based permutation test procedure (100,000 randomizations) to correct p-values [51,52].

2.5. Biological Assays

661W murine cells, which express photoreceptor and retinal ganglion markers [53,54], were used for biological experimentation. We selected this cell line because there is substantial evidence that the 661W cell line can respond to various external stimuli, such as electromagnetic radiation in the visible range, highlighting its versatility in signal reception and transduction (i), and because it is an adherent culture, which could facilitate the transmission of mechanical waves from the microenvironment into the cell (ii). Cells were cultured in DMEM (4.5 g/L glucose) supplemented with penicillin/streptomycin, 2 mM L-glutamine, and 10% (v/v) fetal bovine serum (all reagents were acquired from Capricorn Scientific GmbH, Ebsdorfergrund, Germany). They were seeded at a density of 100,000 cells/flask and incubated in a CO₂ incubator (37 °C and 5% CO₂).

2.5.1. Crystal Violet Staining

The protocol of Mickuviene et al. (2004) was used to evaluate cell proliferation [55]. Cells were seeded in sterile polystyrene 96-well plates (3000 cells/well, 12 replicates) and incubated for 24 h. Afterward, plates were exposed to the acoustic stimulus for 72 h and they were fixed with 96% (v/v) ethanol for 10 min, followed by staining with 0.05% (w/v) crystal violet dissolved in 20% (v/v) ethanol for 30 min. After three consecutive washes with distilled water, the plates were dried in the dark at room temperature. In order to dissolve the dye, 100 μL of a 50% (v/v) ethanol and 0.1% (v/v) acetic acid solution was added and the plate was incubated for 5 min at room temperature. The supernatant was transferred to a clean 96-well plate. Ultimately, the absorbance was measured in a spectrophotometer (AD 340 Microplate Reader, Beckman Coulter, Brea, CA, USA) at a wavelength of 620 nm.

2.5.2. Immunocytochemistry

Immunofluorescence was used to study two conditions, namely (i) the control and (ii) the emotional signal, both of which were incubated separately for 72 h at an equivalent L_p of 80 dB. Sterile round glass coverslips (12 mm diameter) were placed at the bottom of 24-well polystyrene plates. 661W cells were seeded onto the coverslips at a density of 20,000 cells per well, with five replicates per condition. After stimulation, cells were fixed using 4% (w/v) paraformaldehyde (Merck, Darmstadt, Germany) in phosphate-buffered saline (PBS, pH 7.4) for 10 min at room temperature. Then, the fixative solution was removed and the cells were washed three times with PBS (1 mL per well) for 5 min each. Permeabilization was performed for 10 min using 0.1% (v/v) Triton X-100 (Merck) in PBS. To reduce antibody binding to non-specific epitopes, cells were incubated for 1 h at room temperature with 1% (w/v) bovine serum albumin (Merck) in PBS. Next, incubation with the primary antibody against Ki-67 (#28074-1-AP, Proteintech, Rosemont, IL, USA), which was diluted to 1:100 in blocking solution (final concentration 4 μg/mL), was carried out overnight at 4°C in a humidified chamber. Following incubation, cells were washed three times with PBS for 5 min each. A secondary antibody, Alexa Fluor™ 488-conjugated donkey anti-rabbit IgG (#A21206, Invitrogen, Waltham, MA, USA) diluted to 1:1000 in blocking solution (final concentration 2 μg/mL), was then applied and incubated for 2 h at room temperature. Subsequently, cells were washed three times with PBS for 5 min each and coverslips were mounted onto microscopy slides using a DAPI-containing (2 μg/mL) antifade mounting medium (CitiFluor™, Electron Microscopy Sciences, Hatfield, PA, USA). Coverslips were sealed with nail polish and samples were stored at 4 °C, protected from light, until image acquisition. All immunostaining procedures included negative controls, in which the primary antibody was omitted. Imaging and analysis were performed using a laser scanning confocal microscope (LSM 800, Zeiss, Jena, Germany) at 40× magnification.

2.5.3. Image Processing

Confocal images were captured on the same day, maintaining consistent laser intensity, gain, and brightness settings. Parallel staining was performed using identical antibody dilutions. For further image analysis, five random fields per condition were processed. Initially, the DAPI channel images were binarized by adjusting the luminance threshold and designating each nucleus as a region of interest (ROI). The mean luminance value for each cell was then calculated based on the ROI. Luminance assessment was conducted with ImageJ software (version 1.54f).

2.5.4. Statistical Analysis

MATLAB software (version R_2020a) was employed for statistical analysis. Once Grubbs’ test was performed and basic statistics were calculated, the normality of the samples was verified by means of the Kolmogorov–Smirnov test. Variances were compared using a two-sample t-test.

3. Results

3.1. Evaluation of the Affective State

The value of the gratitude letter was calculated from NRC-VAD Lexicon dictionary. When two words matched, the valence and arousal values were extracted from the affective dictionary. To solve the unmatched cases because of the inflected form (i.e., plural, feminine, conjugated, etc.) a lemmatization process was carried out. Then, we calculated the differences between valence and arousal. The gratitude letter valence was 15% more positive with respect to the Don Quixote text, while differences in arousal were less than 1%.

Moving forward, to explore whether there was an alteration in the emotional brain areas during the reading of the gratitude letter, the EEG signal was subjected to analysis. We averaged the log-transformed power spectrum across the 45 windows for each condition and found significant differences between resting-state and gratitude conditions within the range from 1 to 11.125 Hz (Figure 2). We divided and computed the log-transformed power for each canonic frequency and found significant differences between conditions in the theta (p = 0.0020), alpha (p = 0.0257), and beta (p = 0.00257) but not in the gamma (p = 0.5824) band (Figure 3). All p-values were corrected for multiple comparisons (see Methods—Statistical Analyses).

3.2. Audio Analysis

To characterize the acoustic stimulus, the spectrograms, spectra, and MSC were computed. The spectrograms of (ii) the emotional voice and (iv) the synthetized emotional voice were calculated to analyze the similarity of temporal components (Figure 4a). The spectra of these signals were then calculated (Figure 4b). In the context of the MSC, a value of 1 represents the maximum degree of similarity, whereas a value of 0 represents the minimum degree of similarity. Following a visual inspection of the spectra, we calculated the MSC for all the frequencies between 0 and 5 kHz (Figure 4c).

3.3. Characterization of the Environment and Electro-Acoustic Radiation System

The experimental set-up comprised a passive vibro-acoustic treatment and an electro-acoustic radiation system. Initially, the environment was characterized by measuring the L_p both before and after the passive acoustics elements were installed in the CO₂ incubator (Figure 5).

The PVT significantly reduced the background noise from electro-mechanical systems and actuators of the CO₂ incubator, and an overall noise level reduction of 10 dB at 250 Hz, 5 dB at 1000 Hz, and 20 dB at 5 KHz was achieved.

3.4. Effects of Emotional, Non-Emotional and Synthesized AAW on Cell Proliferation

To evaluate the impact of emotions on cellular dynamics, 661W cells were subjected to different AAW stimuli generated by a loudspeaker in a controlled acoustic environment for a period of 72 h at 80 dB. In the control condition, the loudspeaker was inactive. A PVT was installed in the CO₂ incubator test. The analysis of cell proliferation (crystal violet assay) revealed statistically significant differences between (i) the control and (ii) the emotional voice (* p = 0.004), with average absorbances of 0.71 ± 0.02 and 0.80 ± 0.05, respectively. No other significant differences were identified between (iii) the non-emotional voice (p = 0.179) and (iv) the synthesized emotional voice (p = 0.481). The average absorptions were 0.74 ± 0.05 and 0.69 ± 0.05, respectively, in comparison to the control condition (Figure 6c).

As expected, an increase in cell proliferation was accompanied by an increase in Ki-67 expression. The statistical analysis of the confocal images of Ki-67 immunofluorescence (Figure 6a,b) revealed significant differences in cell-to-cell luminance between (i) the control and (ii) the emotional voice (**** p= 0.000007, Figure 6d).

4. Discussion

Given that emotions modify the characteristics of the phonatory apparatus influencing different speech parameters [15,16], and several studies in the field of sonobiology have demonstrated the capacity of AAWs to induce changes in cell behavior [34,35], this investigation sought to identify whether the physical–acoustic changes produced by emotions in speech production could have any effect on cellular dynamics. Among all available emotions, gratitude was chosen because there is an effective method to induce feelings of gratitude in a laboratory setting: gratitude letters [30,31].

At the level of the brain, reading a letter of gratitude written by the same participant following the protocol of Toepfer and Walker produced a change in the participant’s brain activity. While we cannot conclude that changes in EEG power are uniquely due to gratitude, there were significative differences between reading aloud a fragment of Don Quixote and a personal gratitude letter in the theta, alpha, and beta bands. However, the results obtained are in concordance with several studies that reported neural correlates of gratitude in ACC [27,28]. Furthermore, the gratitude letter analysis revealed notable differences in valence between reading aloud a modern version of Don Quixote and a personal gratitude letter. Combining EEG measurements and affective text analysis, we showed that the participant’s emotional system was positively altered when he was in a state of gratitude. The participant reported to the researcher that he felt “calmer and happier” after writing and reading the gratitude letter. As a result, we support Toeper and Walker’s hypothesis that writing a gratitude letter can promote a positive emotional state.

Before performing the biological assays, a PVT was installed in the CO₂ incubator, which provided better acoustic control of the environment and reduced background noise from electro-mechanical systems and actuators. The signal-to-noise ratio was significantly improved, reaching values of 10 dB at 250 Hz, 5 dB at 1000 Hz, and 20 dB at 5 KHz. Biological assays showed a significant increase in the proliferation of 661W cells after exposure to the gratitude voice compared to the control group. This result is consistent with the findings from the crystal violet and the Ki-67 (well-known proliferation marker [56]) immunocytochemistry assays. It is reasonable to assume that the differences shown in 661W cell proliferation between the control and the emotional voice case, a read-aloud version of the gratitude letter, was caused by some particular properties of the participant’s voice (* p = 0.004). However, this possibility was controlled using a non-emotional stimulus, a read-aloud modern version of Don Quixote, which did not produce significant changes compared to the control group (p = 0.179). To isolate the physical aspects of the gratitude emotion in the participant’s voice, a synthesized emotional stimulus was created. This consisted of a synthesized version of the same gratitude letter read aloud by the participant and showed strong temporal synchrony and acceptable spectral similarity (MSC = 0.21, 0–5000 Hz band). The synthesized case was not statistically significantly different from the control (p = 0.481).

As introduced earlier, emotion in speech is mainly characterized by F0, amplitude, timbre, and temporal aspects. Having carefully considered all of these aspects, we hypothesize that the biophysical emotional parameters driving the biological changes may be the F0 and timbre components. The amplitude and temporal aspects showed a high degree of similarity, whereas F0 and timbre showed more variability. Although we cannot rule out that some of these differences may be due to the method of synthesis, variations in F0 are closely linked to emotional expression. For instance, emotions such as happiness often lead to increased F0, whereas tenderness states may lower it [57]. Moreover, the harmonic structure contributes to the timbre of the voice and plays a crucial role in conveying emotional nuance, as different emotional states modulate both the amplitude and distribution of these harmonic frequencies [58,59]. As a future improvement to our protocol, an actor would read aloud the gratitude letter written by the participant, and the resulting AAW signal would be radiated over cells. This new control would provide us with more information about the frequency components.

From a biophysical perspective, there is evidence in the literature that AAW stimuli are capable of inducing significant changes in cellular dynamics and genetic expression. Kumeta et al., 2018 demonstrated that sound stimulation (94 dB, sine-wave sound with a frequency range of 55 Hz to 4 kHz) resulted in a significant reduction in Ptgs2 mRNA levels in the bone marrow-derived ST2 cell line. This was attributed to an increase in RNA decay [34]. In another study by Lin et al. (2021) on in vitro cardiac muscle HL1 cells, significant differences between emotional and non-emotional AAW stimulation were found using alpha-tubulin staining after the different 20 min sound exposures. Compared to the control condition, non-acoustic stimuli, the light emission intensity displayed by phalloidin (F-actin), beta-actin, and alpha-actinin-1 increased when the cells were exposed to meditative music, mantras, and love signals, but decreased with hate signals [60].

Focusing on the mechanotransductional response, the potential biological mechanisms that could trigger the cellular responses to AAW can be classified into two main groups: biochemical signaling pathways and cytoskeletal dynamics. As expected, both groups are interconnected, making it quite difficult in many cases to differentiate between these mechanisms. First, the signal must be captured from the extracellular environment and transmitted into the cell through mechanosensitive sensors. At this point, the plasma membrane contains numerous mechanosensitive biomolecules/structures, notably integrins [61], mechanosensitive ion channels [62], primary cilia [63], cadherins [64], and ICAMs [62,65], among others. When the biosignal reaches these mechanosensors, it induces conformational changes that initiate intracellular signaling. For example, variations in the electrochemical gradient occur via mechanosensitive ion channels [62] and phosphorylation cascades such as the Hippo pathway, which regulates nuclear regulators YAP/TAZ (Yes-associated protein/transcriptional coactivator with PDZ-binding motif) [66]. Additionally, traction forces are generated and transmitted through the integrin–cytoskeleton axis to the nuclear membrane, where various cytoskeletal components physically interact with the LINC complex (linker of nucleoskeleton and cytoskeleton) [67]. Once in the nucleus, different biochemical and/or mechanical signals induce changes in chromatin organization, due, among other factors, to the interaction between the nuclear lamina and the lamina-associated chromatin domains (LADs) [68], creating specific gene expression programs and modulating aspects such as cell differentiation and proliferation [69]. Furthermore, the mechanoregulation of transcription factors through their differential nuclear translocation plays an important role in modulating gene expression. An example of this is YAP/TAZ, which, when phosphorylated through the Hippo pathway, continues to be retained in the cytoplasm, preventing its translocation to the nucleus and its interaction with members of the TEAD family (Transcriptional Enhanced Associate Domain) [70]. However, in its dephosphorylated state, it undergoes nuclear internalization, helping to promote the expression of anti-apoptotic genes such as COX-2, BIRC5, and Glut1 [71].

In summary, although this is a preliminary study that requires an expanded sample size, the hypothesis that emotions could act as biophysical agents influencing cellular processes through the physical and acoustic characteristics of the human voice is a tangible possibility. Further exhaustive research is needed.

5. Conclusions

Emotions impact the characteristics of the phonatory system, influencing physical speech parameters. Additionally, multiple studies in sonobiology have shown that AAWs can induce changes in cellular dynamics. The present work studies whether the physical–acoustic changes induced by gratitude in human speech can influence cell proliferation and Ki67 expression in the 661W cell line (non-auditory). Gratitude, understood as the state of awareness and appreciation that arises when another person’s altruistic actions are directed toward us, was the emotion chosen for the present study due to its social importance. The biological assays demonstrated a notable increase in the proliferation of 661W cells upon exposure to an emotional voice of gratitude compared to the control group. Possible physical and biological pathways were discussed based on the recent literature. Despite the limitations, the possibility that gratitude might influence cellular behavior through the physical–acoustic characteristics of the human voice is a plausible one. Consequently, studying the impact of emotional AAW on cell biology opens up a promising area of research, with the potential to advance our understanding of how human cognition interacts with biology through the laws of physics.