The Scent of Emotion: A Pilot Study on Olfactory Perception Beyond Visual Cues

Abstract

From an evolutionary perspective, smell and taste are the oldest human senses. Despite this, other than chemical senses—particularly vision—are commonly regarded as the most powerful tools for interacting with our environment. Within such a frame, it has become a common belief that blind individuals, especially those who are congenitally blind, develop a compensatory sensory pattern, enhancing the power of their sense of smell. However, the literature results are unclear, mainly due to the heterogeneity of the study population and of the investigation methods. Emotional reactions to olfactory stimuli in blind individuals remain underexplored, primarily due to challenges in delivering stimuli in a standardized and unbiased manner suitable for quantitative assessment. In such a framework, the present pilot study sought to indirectly discover the emotional responses of blind individuals to a specific class of sensory stimuli through the application of wearable sensors for capturing electrocardiographic (ECG) signals and galvanic skin response (GSR). Tonic GSR varied in blind individuals (p < 0.001), but not in controls. Notably, variations were observed between Baseline and Odor 1 (p = 0.002), Odors 1 and 2 (p = 0.003), Odors 2 and 3 (p = 0.003), and on the GSR phasic peak between Baseline and Odor 1 (p = 0.001). No differences were observed for ECG; however, blind individuals’ heart rate correlated with reported pleasantness (r = 0.436, p = 0.005). In light of the different patterns retrieved across stimulus responses, particularly in the GSR signal features, the comparison with a group of non-visually impaired peers shed light on the peculiarities in the psychophysiological responses of blind individuals, with potential use for tailored treatments for the improvement of well-being or, in some cases, for practical applications fostering social inclusion for affected subjects.

1. Introduction

Smell and taste are, for human beings, phylogenetically the oldest of our five senses; they developed much earlier than other types of senses like vision, hearing, and touch [1]. However, it is a common belief that vision is the most powerful sense for interacting with the surrounding environment and, notably, it is considered to be the most important of the five human senses [2,3,4]. Due to this view, it was postulated that people affected by blindness, especially congenital blindness, would develop a sort of compensatory plasticity at the brain level, optimizing the use of brain areas devoted to non-visual processing, thus allowing them to exploit the other senses in an enhanced manner [2,3,4,5]. The reasons for this compensatory effect can be manifold. For example, in sighted individuals, visual regions like the occipital cortex, beyond being active during olfactory identification and discrimination, are mainly activated by visual stimuli, whereas the same cortical portion is activated in blind counterparts, even in the absence of vision, during odor detection, odor discrimination and odor categorization tasks, suggesting an active role of visual brain regions in olfactory processing in blind individuals [3,4,5,6].

Additionally, the absence of visual processing makes blind individuals more attentive to olfactory and environmental information during everyday life, in turn promoting enhanced olfactory abilities [7,8] and possibly modulating the size, and the plasticity, of the olfactory bulb, a key structure for the processing of odors, even if mixed evidence for this is present in the literature [6,9,10,11]. Olfactory attention is deemed a cornerstone for olfactory performances, as it is demonstrated that a conscious focus on odors may change their perception, and that social and physical environment can effectively stimulate the human olfactory system, which is likely to support the improvement of olfactory sensitivity [12].

Nevertheless, due to the aforementioned importance placed on non-chemical senses, it is quite difficult to refer to published studies dealing with olfactory processing in blind individuals, given the paucity of such investigations in the current literature. Similarly, investigations around the objective, quantitative emotional processing of sensory stimuli are still in their infancy in the scientific community in general, due to the scarcity of reliable tools for assessing physiological signals until a few years ago; just recently, thanks to the widespread diffusion of wearable devices, this field of investigation has gained momentum and it is now widely employed to counterbalance the typical biases of sensory analysis, including judgment bias and intra-operator and inter-operator variability, which might affect the reliability of the results obtained. In fact, although explicit measures, typical of sensory analysis, are quick to collect and easy to use, and are well recognized for their capability of assessing the experiential aspects of emotion, implicit measurements are now commonly used alone or in conjunction with explicit measurements in both food science and neuromarketing, with good success in detecting emotional responses to sensory stimuli [13,14,15,16].

However, when focusing specifically on the problem of detecting emotions elicited by sensory stimulation in blind individuals, the current scientific literature is extremely scant. One of the most intriguing studies in this regard was performed years ago by Iversen and collaborators [17], which, based on a cohort of 14 congenitally blind individuals and matched controls, found that blind people displayed an enhanced use of olfaction in assessing the emotional states of other individuals, highlighting their optimization of this sensory channel in recognizing social cues. Research around emotions in visually impaired individuals displayed no significant abnormalities in terms of the atypical or dysfunctional processing of emotions in such population [18,19], although with a potential emotional enhancement in this cohort [20] and, a potential need to promote the use of non-visual emotional signs and body language since early infancy to develop good emotional awareness and emotion regulation [19].

With those premises in mind, we tried to contribute to the current knowledge around this topic by conducting a pilot study dealing with the study of physiological responses to sensory (olfactory and trigeminal) cues in blind individuals and sighted counterparts as an intriguing clue to the emotional responses of these individuals. Regarding sensory triggers, we chose to use spices and aromatic herbs (SHs), based on the previous experience and expertise of our research group in this specific category (see [21,22] for an overview). Spices and aromatic herbs are particularly suitable for sensory research due to their complex and intense olfactory profiles, which, together with the trigeminal contribution, elicit strong emotional and physiological responses. Moreover, SHs have been integral to both cuisine and traditional medicine since ancient times. Unlike vegetables and fruits, which are mainly consumed for their macronutrient content, SHs are used in small amounts to enhance the flavor, color, and presentation of foods. This is made possible by specific chemical compounds that impart distinctive sensory characteristics, including terpenes, alkaloids, flavonoids, phenolic compounds (such as polyphenols), and salicylates. The significance of spices lies in their secondary metabolites, which confer a range of beneficial properties relevant to human health and food technology, contributing to both technological and organoleptic functions. Given the prominent cultural and sensory roles of spices and aromatic herbs, it is reasonable to assume that these botanicals may evoke diverse emotional responses during tasting experiences. Overall, their widespread familiarity and cultural relevance make them ideal candidates for comparative studies across different populations. In the present investigation, all things considered, based on the evidence of enhanced sensory vigilance and cortical cross-modal reorganization in blind individuals, we hypothesized that sensory stimulation would elicit stronger autonomic activation in the blind individuals compared to sighted controls. Specifically, due to the well-recognized role of spices as modulators of emotions in general in the consumer [23,24], and given our past results with spices in sighted individuals [21], we expected higher electrodermal responses and heart rate modulation during stimuli presentation in those individuals used to live with sighting problems, so experiencing heightened sensory attention with respect, for example, to olfactory stimuli, reflecting heightened arousal and engagement with sensory cues.

2. Materials and Methods

2.1. Spices

For the present study, five spices were used as sensory triggers. The compounds were selected due to their peculiar sensory and emotional characteristics, as previously outlined [21]. Details for such samples are reported in

Table 1. Compounds used for sensory stimulation.

Botanical Name	Family	Distribution Area	Common Name	Edible Organs	Production Methods
Piper nigrum L.	Piperaceae	Hainan, Yunnan, and Guangdong in China and Europe	(Eng) black pepper, (Fra) poivre noir, (Esp) pimienta negra, (Deu) schwarzer pfeffer, (Ita) pepe nero	Fruit and bark	Berries are harvested at early ripening when they turn yellow. Then, berries are washed in hot water, and finally, they are sun-dried or dried by artificial methods [25,26]
Piper cubeba L.f	Piperaceae	Sri Lanka, Sumatra, Malaysia, Southern Borneo, and Java	(Eng) cubeb pepper, (Fra) poivre cubèbe, (Esp) pimienta cubeba, (Deu) kubeben pfeffer, (Ita) pepe cubebe	Fruit	The fruits are harvested by hand when ripe and separated from spikes. Then, the berries can be directly dried or immersed in water to remove the pericarp, and afterwards, they are dried for 3–4 days [27]
Piper longum L.	Piperaceae	India, Malaysia, Indonesia, Singapore, Sri Lanka	(Eng) long pepper, (Fra) poivre long, (Esp) pimienta larga, (Deu) langer pfeffer, (Ita) pepe lungo	Dried infructescence and leaves	The infructescence is harvested before ripening when the color is blackish green. Subsequently, the berries are dried in the sun for about 4–5 days [28]
Schinus terebinthifolius Raddi	Anacardiacee	Central and North America, Europe, Asia, and Africa	(Eng) pink pepper or false pepper, (Fra) faux poivrier, (Esp) pimienta de brasil o pimienta rosada, (Deu) rosa pfeffer, (Ita) pepe rosa	Fruit	The berries are harvested manually once they have reached maturity and then dried [29]
Pimenta dioica (L.) Merrill	Myrtaceae	West Indies (Jamaica) and Central America (Cuba, Mexico, Brazil, Honduras, Guatemala, Belize)	(Eng) Jamaica pepper or allspice, (Fra) poivre de la jamaïque, (Esp) pimienta de jamaica, (Deu) jamaika pfeffer, (Ita) pepe garofanato	Fruit and leaves	The harvested berries are left for up to 5 days in sacks to ferment. Then, they are dried for 5 to 10 days, depending on the weather, until the moisture content is about 12% [25]

.

2.2. Study Procedure

All individuals were asked to sit on a chair comfortably during the whole procedure. The testing phase was performed in a clean, well-ventilated room, with constant temperature and humidity (23 °C, 70% R.H.) during the session. All participants were informed about the aims of the study and were asked to sign their written informed consent. In case of their inability to do so, a legal tutor, if present, was asked to sign on their behalf. The protocol was performed under the Helsinki Declaration guidelines, and ethical approval was obtained (CNR Ethical Clearance protocol number 0291041, 21 August 2024). After seat adjustment, the volunteers were individually blinded (in case of sighted individuals) and equipped with wearable sensors for a 1’ baseline recording. After this phase, the five odorous compounds were administered sequentially, following the same order (fixed for all participants) indicated in Table 1, with each compound administered for 10 s and an inter-stimulus duration of 30 s. The stimuli presentation was performed using vials containing the actual compound placed under the nostrils of the participants at a distance of approximately 2.5 cm. The participants were not aware of the content of the vials. The volunteers were asked to “start” and “stop” sniffing by the operator to deal with the actual duration of stimulus presentation. The number of sniffing actions was left under the control of the participant in order to avoid further annoyance and to optimize the participants’ compliance with the experimental protocol. During the inter-stimulus phase, the subjects were asked to rate the pleasantness of the odor just perceived on a 0–9 scale, where “0” represents the most unpleasant and “9” stands for the most pleasant sensation ever.

The physiological signals were then analyzed concerning the baseline and the single stimulus response, as detailed later.

2.3. Psychophysiological Assessment

The whole physiological signal recording process was performed using wearable devices. Two popular signals allowing for the detection of the functioning of the autonomic nervous system (ANS) were captured: (i) electrocardiogram (ECG), displaying the electrical activity of the heart, and (ii) galvanic skin response (GSR), mapping the electrical activity of the skin caused by the activation of sweat glands. Both those signals are reliable indicators of the ANS activity, and the use of wearables in this regard is gaining momentum in scientific research, thanks to their enhanced reliability in performing such measurements [30,31].

2.4. ECG Acquisition and Processing

An ECG signal was acquired using a commercial sensor, the Shimmer2 ECG (Shimmer Sensing, Dublin, Republic of Ireland), capturing the body signal through its adhesion to a commercial fitness chest strap (Polar Electro Oy, Kempele, Finland). The device, equipped with Bluetooth, through which it is capable of communicating with a properly developed acquisition user interface, was programmed to acquire the ECG signal at 500 Hz to comply with the international guidelines for the estimation of the heart rate (HR) and its variability (heart rate variability (HRV)) [32].

The ECG signal was then processed by means of a purposely developed routine in Matlab R2024a (The MathWorks, Inc., Natick, MA, USA). At the beginning, the signal was pre-processed for artifact removal, with QRS complex detection using the Pan–Tompkins algorithm and the RR series reconstruction to correct non-sinusoidal beats. Then, the most important features of the ECG signal, in both time and frequency domains, were extracted, including: heart rate (HR), standard deviation of the normal R-R intervals (SDNN), percentage of normal R-R intervals differing for more than 50 ms (pNN50), Cardiac Sympathetic Index (CSI), Cardiac Vagal Index (CVI), and low-to-high-frequency component ratio (LF/HF) [31,32,33].

2.5. GSR Acquisition and Processing

GSR signals were acquired through a commercial sensor, similar to the previous one, the Shimmer3GSR+ (Shimmer Sensing, Dublin, Republic of Ireland), attached to two adjacent fingers (i.e., the index and middle fingers) of the non-dominant hand via two soft, comfortable rings attached to dry electrodes. Thanks to the Bluetooth connection, the device communicates with a dedicated user interface developed by Shimmer Sensing, where the operator can set the sampling frequency and other parameters.

In the present study, the GSR signal was acquired at 51.2 Hz, as happened in previous research (e.g., [21,33,34]). After the acquisition phase, the GSR signal was processed using Ledalab V3.4.9, a Matlab-based tool specifically developed for GSR processing [35]. The signal was first filtered using a first-order Butterworth low-pass filter at 5 Hz, suitable for the removal of high-frequency noise. Then, a continuous decomposition analysis, which is a robust method for decomposing a GSR signal into continuous tonic and phasic components using standard deconvolution [35], was applied to extract tonic and phasic phases of the signal, and significant features were calculated, including global GSR signal, tonic GSR component, phasic GSR component, and peak value of the phasic GSR component.

2.6. Study Population

A total of 30 participants, aged 44.0 ± 14.6 years (10 males, 20 females), volunteered to enter the study. Of these, 15 were bilaterally blind individuals (age 44.8 ± 15.1 years, age range 20–68 years, 5 males, 10 females) and 15 were sighted individuals (age 43.1 ± 14.5 years, age range 25–65 years, 5 males, 10 females).

Visually impaired participants were recruited thanks to the collaboration with the non-profit institution “Italian Union of the Blind and Partially Sighted” (“Unione Italiana Ciechi ed Ipovedenti”, UICI), Lucca Unit, whereas matched controls were recruited among individuals experienced in sensory analysis, trained sensory panelists at the Department of Agriculture, Food and Environment (DISAAA-a), University of Pisa. The groups did not differ in terms of gender (p = 1) or age (p = 0.867). Blind individuals reporting no residual perception of light, no chemosensory abnormalities, and an absence of neurological factors were included in the study population. All blind participants received a brief training in sensory analysis, conducted by DISAAA-a teachers, which was similar to the one experienced by control panelists, following previously described procedures [36] but limited to olfactory and gustatory modalities due to their clinical condition not allowing the teachers to delve into vision in depth.

2.7. Statistical Analysis

Normal distribution for any of the parameters under investigation was checked using the Shapiro–Wilk test. In the case of normally distributed values, to analyze differences within the same population, a repeated-measures ANOVA followed by Student’s t-test was applied to compare the pair of values, while Friedman’s test followed by the Wilcoxon signed-rank test was used in case of data deviating from normality.

When studying differences between groups (e.g., blind vs. sighted individuals), Student’s t-test was used in case of a normal distribution of data, whereas a non-parametric Mann–Whitney test was applied for data deviating from normality. Correlations between physiological data and pleasantness were studied using Spearman’s correlation test. For all the analyses carried out, the statistical significance was kept at p < 0.05.

3. Results

The perceived pleasantness for the five compounds used, measured on a 1–9 scale, did not differ between the two groups, as outlined in

Table 2. Perceived pleasantness (NS = not significant at p = 0.05 level).

Spices’ Common Name	Blind Individuals (Mean ± SD)	Controls (Mean ± SD)	p-Value
Black pepper	5.46 ± 1.62	6.00 ± 1.41	NS
Cubeb pepper	5.50 ± 1.83	4.86 ± 1.75	NS
Long pepper	6.83 ± 1.64	6.64 ± 1.74	NS
Pink pepper	6.29 ± 1.84	5.93 ± 2.02	NS
Jamaica pepper	5.96 ± 2.16	6.43 ± 2.10	NS

.

At baseline, the two groups significantly differed from each other in the feature LF/HF, with the value referred to blind individuals significantly higher than in controls (1.908 ± 1.381 vs. 0.413 ± 0.485 µS, p = 0.010), probably suggesting higher sympathetic arousal for blind individuals at the beginning of the assessment.

Comparing the same population across different phases, blind individuals reported significant variations throughout the assessment on the Tonic component of the GSR signal (p < 0.001 throughout the test, with notable variations between Baseline and Odor 1—p = 0.002, between Odors 1 and 2—p = 0.003, and between Odors 2 and 3—p = 0.003; see Figure 1). Also, they experienced variations (p < 0.001) in the peak of the GSR phasic phase, particularly between Baseline and Odor 1 (p = 0.001). No significant variations were observed regarding the features related to the ECG signal.

Control individuals displayed significant differences in the GSR phasic peak (p = 0.006), with the most remarkable differences between Baseline and Odor 1 response (p = 0.029). Even in this case, the ECG signal failed to reveal variations throughout the testing phase.

Concerning the correlation between physiological signals and explicitly defined odor hedonics, a significant, moderate, positive relationship was observed between HR and pleasantness (r = 0.436, p = 0.005, Figure 2) just in blind individuals, highlighting that most pleasant odors are associated with an increased sympathetic response in those subjects. No relationship was otherwise observed among sighted counterparts.

4. Discussion

The present investigation was the first, to the best of our knowledge, to assess implicit responses to odors by studying physiological parameters through wearable sensors in blind individuals, with a protocol similar to recent investigations around non-clinical subjects performed by the same research group (e.g., see [21,33,34,36]). The research line around the study of implicit reactions to sensory stimuli is gaining momentum in the scientific literature. In fact, studies using physiological signals like ECG, GSR, or electroencephalogram (EEG) have experienced a growing interest [34,36,37,38,39] due to the mounting usability and reliability of wearable devices, and to the fact that physiological signals are capable of providing useful insights into the physiological response of an individual. Therefore, they can also be used in subjects with important clinical conditions, like dementia, or in very young infants who are still unable to communicate their feelings and emotions. Furthermore, their employment enables a decrease in judgment and cognitive biases, which are among the most important sources of error during sensory analysis or similar frameworks [37].

However, despite their mounting prevalence in scientific literature, such approaches have never been used to study emotional responses to sensory (olfactory) stimulation in blind individuals, especially in comparison with sighted individuals. The problem of chemosensory processing in visually impaired subjects is of paramount relevance in psychophysiology, perceptual sciences, and neurobiology in general. Several works have attempted to compare the functionality of sensory structures in blind individuals with those in sighted controls. A number of studies performed on animals revealed massive structural changes in various brain areas, and not only the visual cortex, among individuals visually deprived at very young ages, with the deprived cortex also becoming responsive to several non-visual inputs [38,39]. Some years later, this phenomenon was also confirmed in human beings, where morphological changes were observed in visually deprived individuals [40], and with the involvement of the visually deprived occipital cortex in processing information from other sensory modalities and further cognitive tasks as well [41,42]. Stemming from these discoveries, later evidence discovered a significant occipital activation during odor discrimination and characterization, with the right fusiform gyrus being the most stimulated by olfactory cues, resulting in overall superior performance in odor recognition tests by blind individuals [43,44]. The excellent olfactory performances of blind individuals can be due to both improved perceptual abilities and better access to the information stored in semantic memory since their early days [7]. Interestingly, such an ability was mainly observed for orthonasal performances rather than retronasal abilities, which are in turn related to taste processing, a task that blind individuals appear to find more difficult due to a decreased semantic knowledge of sensory stimuli in this sense [8].

However, beyond morphological, functional, and perceptual processing of olfactory stimuli, no emotional counterparts have been investigated for blind individuals in relation to the administration of odors. Emotional processing for blind individuals is a debated field in scientific research; however, no specific abnormalities have been highlighted for this population in the majority of cases [8,17,18]. Nevertheless, the emotional processing of chemosensory stimuli could be one field of interest for this specific population due to the impact that chemical senses have on some basic functionalities of the human being, including the attitude towards feeding.

Within such a framework, this pilot study involved 30 gender- and age-matched individuals, equally divided into blind and sighted groups, to whom specific odors (i.e., spicy compounds) were administered while physiological signals were recorded.

Concerning the methodology used in terms of signals studied, although all physiological parameters measured are related to ANS activation, they capture different components and temporal dynamics of autonomic regulation. As such, the GSR primarily reflects sympathetic activity associated with short-latency, phasic arousal, while HR and HRV indexes integrate both sympathetic and parasympathetic influences over longer time scales [9,45].

According to our results, blind subjects appeared to be more highly aroused before the test administration, and the variations experienced throughout the sniffing of the five different compounds were more significant. At first, blind participants exhibited higher autonomic arousal before stimulus presentation, possibly reflecting enhanced sensory vigilance and anticipatory attention associated with the compensatory use of olfaction in the absence of vision [7,16]. This appears to be consistent with a structural and functional cortical remodeling among blind individuals, with enhanced efficiency and sensitivity across other sensory channels, including auditory and olfactory ones. Second, the larger fluctuations in ANS responses across different odorants suggest heightened physiological discrimination between different odors for blind individuals, consistent with previous evidence of superior olfactory discriminatory ability in this cohort [46], and for the first time reflected here also in their autonomic counterpart.

In addition, the “blind” cohort members reported higher sympathetic arousal for the compounds judged as more pleasant, confirming a sort of valence-dependent response of the ANS, suggesting that pleasant olfactory experiences evoke robust physiological engagement in the blind, possibly linked to emotions and likely mediated by cross-modal cortical reorganization [16]. In this context, the selective modulation of GSR but not HRV suggests that sensory valence in blind individuals primarily triggers rapid sympathetic arousal rather than sustained autonomic balance shifts. Our explanation for our result is therefore that higher functional coupling between sensory (olfactory) and limbic systems in blinded volunteers is due to cross-modal plasticity. In fact, in the absence of visual inputs, other sensory channels, including olfaction, gain relevance, possibly leading to a stronger autonomic modulation by odor valence. In addition, reduced top-down cortical inhibition on autonomic responses observed in blind individuals may contribute to a more direct physiological response to emotional appraisal [16,38]. The specific contribution of the trigeminal stimulation to the whole response needs to be further elucidated, since, according to the literature, blinded individuals appear not to have an altered response to trigeminal stimuli [47] and, since this type of response appears to be strongly interrelated to the olfactory response [48], further specific investigations need to be performed.

5. Conclusions

The present pilot investigation attempted to shed light on physiological processing, contributing to the emotional response, during sensory stimulation in blind individuals. Despite the relatively low sample size, the lack of randomization in the order of the stimuli proposed—which could eventually account for order effects—and the cross-sectional nature of the investigation, some peculiarities emerged, mapping at the psychophysiological level what had already emerged at a psychophysical degree in past studies, notably an enhanced sensory vigilance, a compensatory use of olfaction in the absence of vision, and heightened physiological discrimination among blind individuals that also reported higher sympathetic arousal for more pleasant compounds, probably mapping an overall reorganization at cortical level. However, the preliminary nature of such findings needs to be mapped with future research to enhance their generalizability.

Starting from such assumptions, future studies, using wearable sensors, should be undertaken on larger, more stratified cohorts, possibly split by the degree of visual impairment and eventually the congenital nature of blindness, to understand whether our findings can be applied to the whole population of blind individuals or whether any difference between congenital blindness and acquired blindness does exist in sensory and related emotional processing. Also, they should use repeated stimulation in order to reduce the likelihood of finding differences by chance, and the usage of arousal explicit measurements should be added beyond valence. Finally, the specific contribution of trigeminal stimulation needs to be better elucidated in this population if we aim to continue using spices as sensory triggers to this end. Overall, beyond its scientific value, the present work may have practical implications. The integration of wearable sensing technologies with olfactory stimulation (well-grounded in other experimental domains) can be useful for designing personalized sensory training and rehabilitation strategies in blind individuals to balance their emotional responses. In addition, these findings may lead to the development of (multi-)sensory stimulation environments to support social interaction and spatial orientation, thereby fostering inclusion and autonomy in visually impaired populations.