Introduction

While several reviews exist on the role of touch in social communication (see Croy et al., 2022; Fairhurst et al., 2022; Gallace & Spence, 2010; Hertenstein et al., 2006b), there appears to be no attempt to categorize all the classes of behavior underpinned by the tactile sense. A broader review of behavioral function in touch is beneficial for several reasons. Focus on a single class of behaviour limits understanding of the versatility of the tactile system in non-verbal communication. Determining behavioral function in touch also provides the basis for understanding tactile perception, an estimation of its value in the face of loss and highlights novel research directions for the study of non-verbal behaviour (Gallace & Spence, 2010). As such, the aim of this manuscript is to categorize the behavioral functions of the human tactile system.

Previous reviews have used a variety of definitions of the tactile system, focusing on: (1) agency (i.e., active [touching something] vs. passive [being touched by something]; Gallace & Spence, 2010; Schirmer et al., 2022); (2) afferent type (i.e., discriminative [object-based touch; involving myelinated afferents] vs. affective [interpersonal touch, involving unmyelinated afferents on hairy skin; McGlone et al., 2007, 2014]); (3) interpersonal function (Olausson et al., 2010; Schirmer et al., 2023) and (4) sub-modality (thermal, itch, texture, pain; Abraira & Ginty, 2013; Gallace & Spence, 2010). To provide the broadest definition of touch, we include any percept arising from contact with and to the body surface.

To develop a theory about tactile function from a behavioral perspective, our approach to searching the literature for relevant material involved four stages. First, we examined past behavioral reviews of touch, screening databases such as PsycInfo, using relevant search terms, such as ‘functions’ and ‘tactual/tactile perception’. All relevant behavioral reviews indicated the relevance of touch to social communication (e.g., Dunbar, 2010). Second, we examined seminal biological reviews of the somatosensory system identified from the search above (e.g., McGlone & Reilly, 2010; McGlone et al., 2014) – as perception may provide information about behavioral function of a modality. Third, we examined integrative reviews of behavioral function in other modalities, as functions identified here may be relevant to touch (e.g., Croy et al., 2014; Stevenson, 2010a) – this time using the search terms ‘functions’ and ‘olfactory/visual/auditory/taste perception.’ In addition, we examined citing and cited papers in identified reviews for relevant studies. Finally, as all authors have researched the behavioral functions of human touch, we drew from our past knowledge of the tactile literature. After adopting this approach, three logical classes of functions emerged, pertinent to an individual’s and species’ survival – i.e., Ingestive Behavior, Environmental Hazard Detection and Management, and Social Communication.

As such, this review is organized around the role of touch in Ingestive Behavior, Environmental Hazard Detection and Management and Social Communication – with evidence from animal, human capacity and human impaired (i.e., individuals with reduced/heightened tactile function) data provided for each identified function.

Ingestive Behavior

The role of the tactile sense in ingestive behavior can be structured around the different stages of eating, starting with nipple localization (for neonates only), bringing food to the mouth, ingestion and expectancy violations. Each of these is turned to below and summarised in Fig. 1.

Fig. 1
figure 1

An integrative overview of tactile functions

Full size image

Nipple Localization in Neonates

The tactile system plays a key role in localization of the infant’s mouth to the caregiver’s nipple during breast-feeding. The neonate’s use of touch for feeding may form the basis for its use in more affiliative functions (mother infant attachment), as turned to in the Social Communication section.

Animal Data

Rats, kittens and rabbits use tactile cues for localization and attachment to the nipple. For instance, Larson and Stein (1984) placed either local anesthetic on the paws and lips of kittens (to block tactile sensations) or intranasal nose plugs (to block olfaction). Kittens could not locate and attach to their mother’s nipple when tactile sensations were blocked but could orient towards their mother. The reverse pattern was found when olfactory sensations were blocked. Similar findings emerge in rabbits (Distel & Hudson, 1985), and rats (Hofer et al., 1981). Taken together, normative tactile functioning seems important for infant feeding behaviors in mammals.

Human Capacity Data

Immediately after birth, skin to skin contact with the mother and infant is found to increase the likelihood of exclusive breast-feeding (Gouchon et al., 2010; Mahmood et al., 2011; Marín Gabriel et al., 2010; Moore & Anderson, 2007a, 2007b; Robiquet et al., 2016; Widström et al., 2019). Skin to skin contact (aka. kangaroo mother care) enhances effectiveness of breast-feeding and reduces mortality in preterm (i.e., low birth weight) infants (see Boundy et al., 2016 for a review; World Health Organisation., 2003). The influence of skin to skin contact on breast-feeding efficacy may be most pronounced two hours post birth, as infants are most receptive to tactile, thermal and odor cues from their mothers during this period (Bystrova et al., 2009; Moore & Anderson, 2007a, 2007b). Yet, the extent to which tactile cues influence breast-feeding efficacy, relative to other modalities during this period, remains unclear.

Human Impaired Data

In the mother, dermatologic conditions on the breast/nipple, which cause increased tactile sensitivity, pain, and inflammation, are related to discomfort during and cessation of breast-feeding (Higgins et al., 2013; Waldman et al., 2019). Thus, normative tactile functioning in mothers appears important for breast-feeding, with more research needed on the relation between tactile sensitivity and breast-feeding efficacy.

Prior to Ingestion: Hand to Mouth

Animal Data

The tactile sensitivity of anthropoid hands is very acute and may have evolved for food handling and edibility assessments (Dominy et al., 2001). In non-human primates, fruits and plants are manipulated by the digits, lips, incisors and feet (i.e., areas with high mechanical textural sensitivity). This manipulation likely provides the primate an initial evaluation of the fruits’ and plants’ softness – i.e., an indicator of ripeness and thus edibility – and toughness – i.e., an indicator of high fiber content and thus inedibility (Dominy, 2004; Lucas et al., 2012; Wrangham, 1975). For example, prior to ingestion, several species of frugivorous primates are found to palpate fruits using their hands to evaluate softness, size, shape and the taste properties of fruit (Dominy et al., 2001; 2004; 2016; Kinzey & Norconk, 1990; Laska et al., 2007; Melin et al., 2022; Pablo-Rodríguez et al., 2015; Sánchez-Solano et al., 2022). Softness is likely a more accurate and prolific indicator of ripeness than color, as softening of fruit progresses faster than pigment alterations (Dominy, 2004; Pablo-Rodríguez et al., 2015), and several fruits are cryptic (i.e., color invariant – e.g., green all year round; Dominy et al., 2016; Sánchez-Solano et al., 2022). In line with this, touch sensitivity (mechanoreceptor density) and fruit consumption were positively correlated in nine frugivorous non-human primate species (Hoffmann et al., 2004). When fruit is rare, fibrous foods (terrestrial piths and leaves) are consumed by non-human primates (Wrangham et al., 1991). Mechanical force felt by the hand and/or mouth (i.e., via exploratory incisor bites, prior to ingestion) likely forms the basis for fiber detection, and in turn the food’s rejection or acceptance, as fiber is colorless, tasteless and odorless (Lucas et al., 2012). Thus, the tactile sense appears pertinent for a primate’s evaluation of a food’s edibility, prior to consumption.

Human Capacity Data

Several somatosensory (tactile) features are used to describe the hand and mouthfeel of foods. These include mechanical (lumpiness, firmness, brittleness, adhesiveness, viscosity), geometrical (particle size and shape; graininess), chemical (i.e., wetness, oiliness) and thermal (warmth) features (Engelen, et al., 2005; Lawless & Heymann, 2010; Stevenson, 2010b; Szczesniak, 1991). Firmness seems to be the most researched tactile quality and for some foods provides an indicator of freshness (Heenan et al., 2009; Jaeger et al., 1998). Humans can accurately discriminate the firmness of solid foods, by palpating them between their fingers and thumbs and bending them – and of liquid foods, by evaluating their viscosity (e.g., watching products fall from a spoon; Szczesniak & Bourne, 1969).

Visual texture can also create expectations about how satiating a food may be (Forde et al., 2013b; Hogenkamp et al., 2010; McCrickerd & Forde, 2016; Pisqueras-Fiszman et al., 2011; Stribiţcaia et al., 2020; Yeomans & Chambers, 2011). Increasing the visually perceived thickness (Stribiţcaia et al., 2022), as well as auditory and oral viscosity (from thin to thick) – increases the expected satiation value of the food (Pellegrino et al., 2019). Thus, before ingestion, hand felt and visual texture can indicate how edible (fresh, ripe) and satiating a food may be.

Human Impaired Data

In children, heightened tactile sensitivity (i.e., sensitivity to non-invasive pressure; Hunt et al., 2017) is related to increased rejection of edible foods with complex textures (fruits, vegetables, proteins), and a higher intake of foods with simple textures (simple carbohydrates – e.g., mash potato, white bread and pasta). This restricted diet causes nutritional deficiencies later in life (Dovey et al., 2008; Taylor et al., 2015), and is associated with various developmental disorders – i.e., avoidant restrictive food intake disorder (ARFID) and autism spectrum disorder (Coulthard & Sahota, 2016; Coulthard & Thakker, 2015; Coulthard et al., 2022; Harris, 2009; Mayes & Zickgraf, 2019; Nederkoorn et al., 2015).

As age increases, dexterity and hand grip capacity reduces (Laguna & Chen, 2016; Thornbury & Mistretta, 1981). Individuals with reduced hand grip strength have greater impediments in eating and reductions in their appetite (Chang et al., 2020; Pilgrim et al., 2015). Poor tactile dexterity may also increase injuries during eating, as packaged food cannot be opened and so inappropriate tools (e.g., knives) are used (Caner & Pascall, 2010; Ford et al., 2016). Consistent with this, there are an estimated 50,000 to 94,000 packaging accidents each year, costing up to 12 million Euros (Lewis et al., 2007). Recent survey data suggests 55% of people aged 65 + , would pay more for less packaging (Adams, 2021). Thus, prior to ingestion, normal tactile functioning appears pertinent to gauging edibility of foods and bringing them safely to the mouth.

Second Stage of Ingestion: Food in the Mouth

Animal Data

Relative to solid foods, soft foods seem to increase food intake, weight gain and result in adverse metabolic changes in rats, but these effects may be dependent on the type of soft-feed diet used (Anegawa et al., 2015; Desmarchelier et al., 2013; Ford, 1977; Han et al., 2018; Oka et al., 2003). Softer foods are probably preferred to harder foods because they are easier and faster to eat and digest, which may then increase preference via associative learning.

Human Capacity Data

Chewiness and hardness in solid foods and thickness, viscosity, creaminess in liquid and semi-solid foods, are associated with reduced food intake and increased satiety (Stribiţcaia et al., 2020). For instance, individuals feel equally full after consumption of harder foods, relative to softer foods, despite a 10–30% reduction in intake (Bolhuis et al., 2014; Forde et al., 2013b; McCrickerd & Forde, 2016; Zhu et al., 2013; Zijlstra et al., 2008). This effect appears independent of food palatability (Bolhuis et al., 2014; McCrickerd & Forde, 2016), and may be because harder and chewier foods, as well as creamier and thicker drinks, have greater orosensory exposure – i.e., require more chewing and smaller gulps or bites (Forde et al., 2013a; Haber et al., 1977; Mattes, 2005; McCrickerd & Forde, 2016; Viskaal-van Dongen et al., 2011). At a cognitive level, foods and drinks that are solid and thicker may be perceived as more satiating than soft or low viscosity foods and drinks, due to associative learning (e.g., Yeomans & Chambers, 2011). In line with this, harder, chewier, thicker and creamier foods and beverages tend to be more nutrient rich, and contain more protein, carbohydrates and fibers than low viscosity foods and beverages (Viskaal-van Dongen et al., 2011).

Human Impaired Data

People with impaired ability to masticate and orally process foods (i.e., ~ 15–29% of patients in long term care settings) are provided with texture modified diets (i.e., foods that are pureed to ensure safe swallowing and consumption; Cormier et al., 1994; Hotaling, 1992). These diets are often considered unpalatable and may lead to reductions in food intake and quality of life (Macqueen et al., 2003; Raheem et al., 2021; Swan et al., 2015; Yver et al., 2018). For example, individuals on a texture modified diet ate 3877 kJ, whilst controls ate 6115 kJ (Wright et al., 2005). Thus, while softer foods are typically preferred, when food is not usually soft, altering its texture profile may facilitate rejection due to violations of its expected flavor.

Expectancy Violation

Animal Data

Primate researchers and zookeepers report that chimpanzees (Pan troglodytes) place food on the lip, visually inspect, taste, and then expel it (Case et al., 2020). This suggests that when expected flavor (based on mouth contact and sight) are incongruent with actual flavor (based on taste, smell and touch in the mouth), the food is rejected.

Human Capacity Data

Incongruency between expected and actual food texture, can alter edibility evaluations of food. Zampini and Spence (2004) asked participants to rate how fresh a chip tasted, whilst wearing headphones, which attenuated the sound (crispiness) of the chips by 0 dB (no attenuation), 20 dB, or 40 dB. Chips were rated as more stale and softer when presented with greater attenuation.

Human Impaired Data

Elderly people with mastication issues report visual appearance (e.g., serving shape, size, color) and the taste of texture modified foods, as most pertinent to their edibility decisions (Rusu et al., 2020). Thus, when the tactile system is impaired, expectations of a food’s edibility are based off properties from the other sensory modalities.

Environmental Hazards

In any environment, a range of things can penetrate the skin’s surface and enter the body. Non-verbal behaviors critical to survival include detection, avoidance and removal of these threats from the skin, as they can pose a great risk of harm (e.g., blood loss, infection, reduced sexual fitness and death; Hart, 1990; Hunt et al., 2017). While the skin serves as an extensive physical barrier to these hazards (Hunt et al., 2017; Ludriksone et al., 2014; McGlone & Reilly, 2010; Nguyen & Soulika, 2019; Proksch et al., 2008), the role of the tactile sense in identification and removal of skin based hazards has not been addressed before.

Two classes of hazards instigate a response from the tactile system: (1) disease related (i.e., parasitic, microbial, infectious) and (2) non disease related (i.e., mechanical, thermal, chemical) hazards. Each class of hazards relates to a different emotional state, respectively disgust and fear (Royzman & Sabini, 2001). Touch is clearly integral to the detection of non-disease related hazards, as demonstrated by rare cases of individuals with early onset loss of touch sensation. These individuals are reported to die in childhood, as they fail to feel and in in turn treat their injuries (see Nagasako et al., 2003). Yet, as little research has examined how we use to touch to detect and respond behaviorally to non-disease (fear) related hazards, this section focusses on the role of touch in detection and removal of disease related hazards only (see Fig. 1 for a visualisation of identified sub-functions).

Detection of Disease Related Hazards

Animal Data

There are four ways animals use the tactile system to detect and prevent disease related threats from attaching to the skin: (1) recognition of tactile cues indicative of disease and pathogen presence; (2) programmed grooming to reduce likelihood of parasite contact; (3) tactile related bodily movements, to reduce the potential of parasites from invading the skin; and (4) tool use to control biting parasites. Each is reviewed below.

Non-human primates can recognize tactile cues indicative of microbial presence (Sarabian et al., 2017, 2018). Sarabian et al. (2017) found that chimpanzees were less likely to eat a banana that was placed on wet dough (i.e., a substrate with moisture, softness, and thermal properties resembling living matter and decay), than a control object (dry rope). Further, bonobos (Pan paniscus) were less likely to consume banana slices that were directly on or in close proximity to feces, than banana slices that were placed at greater distances from the feces (Sarabian et al., 2018). Thus, prior to contacting an object, inspection of its visual texture and proximity to microbial hazards can facilitate avoidance.

Ectoparasites, which feed on their host’s blood and serve as disease vectors, are prevented from latching onto the skin of several wild and captive mammals using a tactile behavior known as programmed grooming (Chapman et al., 2005; Hawlena et al., 2008; Kupfer & Fessler, 2018; Mooring et al., 2004, 2006). Programmed grooming may be centrally driven rather than stimulus driven. Namely, even when there is no tactile sensation in rodents, via amputation of their limbs (Fentress, 1973, 1988) or sectioning of the trigeminal nerve to prevent facial tactile sensation, grooming tendencies are unaffected (Berridge & Fentress, 1987).

Several tactile related bodily movements also reduce the ability of ectoparasites from lodging onto the skin. Cattle and red deer switch their tails, head toss, twitch their ears and foot stamp, which reduce the potential of flies from biting their skin (Harris et al., 1987; Hart, 1990; Woollard & Bullock, 1987). Such fly switching behaviors are so effective that, relative to heifers with intact tails, heifers with docked tails have up to three times as many flies landing on them (Ladewig & Matthews, 1992).

Finally, some animals manipulate objects to facilitate parasite avoidance. For instance, Asian elephants have been shown to grasp and swing branches and vegetation against their body, so to prevent and dislodge biting flies (Hart & Hart, 1994).

Human Capacity Data

There are two ways humans use their tactile system to detect and prevent disease related threats from entering the body: (1) recognition of tactile cues indicative of disease and pathogen presence; and (2) skin guarding behaviors to reduce potential contact with parasite and skin-biting insects.

Humans can detect wetness, stickiness and softness (i.e., textures associated with contamination risk and microbial presence) using visual images alone (Cavdan et al., 2021; Iosifyan & Korolkova, 2019; Lee et al., 2019; Sawayama et al., 2017). Visual texture is likely a reliable indicator of actual texture, because of learnt associations and overlapping neural profiles (i.e., both activate the secondary somatosensory cortex; Sun et al., 2016). Visual cues of wetness can elicit disgust in humans, indicating their role in disease threat detection (Stevenson et al., 2019; Saluja and Stevenson, 2022).

In addition, images of ectoparasites (e.g., maggots) and skin-biting insects (e.g., spiders) can pre-emptively trigger skin crawling sensations (colloquially referred to as the heebie jeebies; Blake et al., 2017; Kupfer & Fessler, 2018). A lecture about parasites (e.g., tapeworm, and roundworms) increased students’ degree of self-grooming behavior (scratching, handwashing), relative to a control lecture (i.e., on hormones; Prokop et al., 2014). Relatedly, individuals report feeling heebie jeebies and disgust sensations when exposed to parasites, skin-biting insects (e.g., spiders, ants) and visual textures, which resemble skin infections (irregular clusters, perforation; Blake et al., 2017; Kupfer et al., 2021).

Finally, tactile sensitivity may increase following exposure to disease related hazards. Hunt et al. (2017) found reliable increases in participants’ tactile sensitivity after exposing them to live maggots. Presumably, increased tactile sensitivity may facilitate faster detection and bodily localisation of disease threats.

Human Impaired Data

People with skin picking or excoriation disorder (~ 1.4 – 5.4% of the population; Hayes et al., 2009; Keuthen et al., 2010) engage in recurrent picking of healthy skin and skin irregularities. These individuals report more disgust and urge to pick their skin when looking at skin irregularities than controls (Anderson & Clarke, 2019; Schienle et al., 2018). Skin picking behaviors in these individuals may be explained, in part, by greater sensitivity to tactile sensations (Houghton et al., 2019). Thus, overactive tactile sensitivity may facilitate inaccurate perception of disease related threats on the skin and pathological surface guarding (skin picking) behaviors.

Removal of Disease Related Hazards From the Skin

Animal Data

After an insect or spider makes contact and bites an animal’s skin, stimulus driven grooming aids in dislodgement of biting parasites from the skin (Hart, 1990). For example, amputation of the limbs in mice significantly increased lice infestation, as the mice were inhibited from grooming in the presence of parasites (Bell et al., 1962; Clifford et al., 1967). Several animals (e.g., deer, rodents) are found to engage in increased grooming behavior, as ectoparasite density increases (Hart, 1990; Hawlena et al., 2008; Heine et al., 2017).

Human Capacity Data

Hand felt textures associated with microbial presence (wet, sticky, soft, and to a lesser extent oily) appear to elicit disgust and may also motivate disease avoidance behaviors (Oum et al., 2011; Saluja & Stevenson, 2019; Saluja et al., 2022; Saluja and Stevenson, 2022). For example, Saluja and Stevenson (2019) presented 9 objects (all varying in texture properties) to blind-folded participants, and found hand felt stickiness, wetness, softness, and unexpectedly coldness, were indicative of disgust. The relationship between coldness and disgust likely occurred because wetter objects feel colder (Filingeri et al., 2014). Objects perceived as disgusting are also touched for shorter durations than those perceived as neutral or pleasant (Saluja et al., 2023). Thus, the tactile sense can detect textures associated with pathogens and their recognition instigates disease avoidant reactions.

Human Impaired Data

Patients with diabetes mellitus have reduced tactile sensitivity and neuropathy, usually at their lower extremities, and are also prone to skin ulcers at these regions (Picconi et al., 2022; Ndip et al., 2012; Singer & Clark, 1999; Targino et al., 2016). Diabetic ulcers often become infected because they are not felt by the individual and in turn, left exposed to environmental hazards (Said, 2007; Singer & Clark, 1999). Thus, loss of tactile sensation impairs the ability to detect skin openings and in turn treat and prevent infections, highlighting its centrality to hazard detection and management.

Social Communication

Based on subsequent empirical data and reviews, we identified three sub-functions of the tactile sense in social communication: (1) affective communication and negative affect regulation; (2) relationship status signaling and (3) reproductive related functions (see Croy et al., 2022; Fairhurst et al., 2022; Gallace & Spence, 2010; Hertenstein, 2002; 2006b for reviews). The evidence for each of these sub-functions is turned to below and summarised visually in Fig. 1.

Affective Communication and Negative Affect Regulation

Affective State Communication

Animal Data

Tactile behaviors may be used to signal positive (i.e., greeting and play) and negative affective states (aggression) in mammals (de Waal, 1989; Goodall, 1968; Nishida, 1970; Wilson & Kleiman, 1974). Head-to-head greetings as well as head-to-chest greetings and huddling have been observed in wild and captive woolly monkeys (Lagothrix; Defler & Stevenson, 2014). In chimpanzees, Goodall (1968) observed that greetings include touching of the hand, shoulders, back, groin, thigh or genital area of conspecifics, and these were contingent on the length of time the primates had been separated, mutual attraction and mood (Hiraiwa-Hasegawa, 1989). Further, allogrooming, may evoke positive affect and maintain social bonds, and facilitate physiological changes synonymous with relaxation in mammals (Aureli et al., 1999; Boccia et al., 1989; Reinhardt et al., 1986; Sato et al., 1993; Val-Laillet et al., 2009). For negative affect, Macaques and Old World monkeys may physically mount each other to communicate aggression and their intention to engage in or avoid a fight (Maestripieri, 1997).

Human Capacity Data

Humans can communicate distinctive emotions using touch. Hertenstein et al. (2006a) asked participants to communicate different emotions on another participant’s arm using touch alone, whilst being videorecorded. Naïve participants were shown these recordings and correctly identified anger (44%), fear (38%), happiness (61%), disgust (71%), love (60%), and sympathy (53%) at above chance levels. Emotion recognition via touch was comparable to vision (facial displays) and sound (vocalizations; i.e., 40–80%; Elfenbein & Ambady, 2002).

Communication of affect via touch appears better between acquainted individuals than strangers. Thompson and Hampton (2011), adopting Hertenstein et al.’s (2006a) paradigm, found that couples were significantly more accurate than strangers at decoding self-focused emotions (i.e., embarrassment, envy and pride). More recently, McIntyre et al. (2022) showed that tactile behaviors used by acquainted individuals (partners, close friends) to communicate emotions can be trained to unacquainted individuals – and doing so improves encoding and decoding of affective states between strangers. Thus, acquainted individuals may develop a specialized touch language to code certain emotions to each other.

More broadly, positive affect and intentions can also be communicated via touch. Slow interpersonal stroking is perceived as soothing and relaxing in humans and higher on arousal/love and intimacy intentions, whilst fast interpersonal stroking is perceived as higher on fear and threat-signalling intentions. The effect of touch velocity on perceived intentions is, in part, because, gentle and caressing touch at ~ 1 to 10 cm/s is sub-served by C-low threshold mechanoreceptors (C-LTMRs) – which innervate the hairy skin and are involved in the perception of affective or social touch (see McGlone et al., 2014, Sailer & Leknes, 2022; Saarinen et al., 2021; Schirmer et al., 2023; Morrison et al., 2010; Olausson et al., 2010, for reviews and seminal papers).

Human Impaired Data

Reduction in pleasantness of affective touch on hairy skin, and/or tactile sensitivity has been associated with reduced emotional awareness (e.g., alexthymia) in both clinical (e.g., anorexia nervosa; Crucianelli et al., 2021) and nonclinical samples (Cazzato et al., 2021). Similarly, autism spectrum disorder is characterized by a lowered capability to gauge and communicate emotions using non-verbal mediums (facially and vocally), and this may be related to heightened tactile sensitivity and/or awareness of affective touch (Croy et al., 2016; Hobson, 1986; Perini et al., 2021). Thus, impaired tactile sensitivity may reduce an individual’s ability to signal and decode differing affective states.

Negative Affect Regulation

Animal Data

In non-human primates allogrooming has been shown to relieve tension caused from aggressive interactions (Aureli et al., 1999; Boccia et al., 1989; de Waal, 1989; Hertenstein et al., 2006b). Stroking can also signal the termination of social interactions in non-human primates (Boccia et al., 1989). In rodents, touch has significant impacts on a pup’s response to, and ability to regulate, their stress in the short and long term (Hofer, 2006). For example, Walker et al. (2022) either exposed rats to mild and chronic stressors for two weeks or left them unexposed (serving as a control group). Before exposure to the stressors, rats received either no stroking, stroking at an affective touch velocity (i.e., 5 cm/s) or stroking at a faster velocity (i.e., 30 cm/s). Rats who received stroking at an affective touch velocity had significantly fewer physiological (i.e., corticosterone levels) and behavioral (i.e., climbing during the forced swim test) responses associated with stress – and their responses in these measures were comparable to rats in the control group. Stroking at either velocity increased the time stressed rats spent the center zone of an open chamber, suggestive of an anxiolytic effect of touch compared to no touch (Walker et al., 2022). Thus, in mammals, touch - especially stroking at an affective touch velocity - assists in regulating negative affect.

Human Capacity Data

Evidence from studies using massage therapy indicates that touch is pertinent in reducing anxiety and regulating negative affect following life stressors (Field, 1998; Field et al., 1996a). For example, in a cohort of children displaying deviant behaviors, children who received daily massage therapy had fewer parent reported aggressive behaviors and social issues, relative to their baseline and to a group of children who received stories from staff members (von Knorring et al., 2008). The role of massage therapy in self-reported negative affect regulation has also been shown in adolescents (e.g., Diego et al., 2002) and adults (e.g., Field et al., 1997; Ironson et al., 1996). In few studies, massage therapy is shown to reduce anxiety, depression and sleep disturbances to a greater extent than other forms of relaxation therapy (i.e., viewing relaxing videos), and to mere physical contact (Field et al., 1992, 1996b). However, few studies also find comparable effects between massage therapy and other forms of relaxation therapy (e.g., Field et al., 1997; Gonçalves et al., 2017).

Tactile behaviors may also be used to regulate negative affect. Following competitive tasks (e.g., a sport’s match, Benenson & Wrangham, 2016; or a test of real world knowledge, Benenson et al., 2018), individuals (especially males) increase their physical contact and proximity to repair relationships with their friends (of the same sex). Thus, the tactile sense seems to play a critical role in reducing negative affect following anxiety, depression inducing, or conflict related events.

Human Impaired Data

Several disorders (i.e., autism spectrum disorder, anorexia nervosa, avoidant restrictive food intake disorder, attention deficit hyperactivity disorder) may be associated with both impaired tactile functioning and emotion regulation difficulties (Harrison et al., 2009; He et al., 2021; Krom et al., 2019; Lecavalier, 2006; Mazefsky et al., 2013; Nederkoorn et al., 2015; Nicely et al., 2014; Steinberg & Drabick, 2015). Despite this overlap, no study appears to have examined whether impaired tactile functioning hinders negative affect regulation.

Relationship Status Signaling

Animal Data

The tactile sense plays a key role across species in signaling relationship status and dominance. Allogrooming of inaccessible body areas is used to maintain social organization and dominance structure by several mammal species (de Waal, 1989; di Bitetti, 1997; Dunbar, 2010; Hertenstein et al., 2006b). In non-human primates, equity of grooming between the groomer and recipient is mediated by few variables, including; (1) linearity of the species social group – i.e., in steep hierarchies, allogrooming equity may, in part, be predicted by social rank and dominance (e.g., Franz, 1999; Schino, 2001; Snyder-Mackler et al., 2016; Tiddi et al., 2012), while in more shallow hierarchies allogrooming may be less affected by social rank and is more equitable (Kaburu & Newton-Fisher, 2015); (2) kinship grooming, which is more reciprocal and beneficial (de Waal, 1989; Schino, 2001); and (3) sex, where conspecifics of similar rank and/or sex (e.g., adult female bonobos) may show equitable allogrooming to stabilize social groups and enhance group cohesion (Franz, 1999).

Increasing proximity and huddling conspecifics has also been observed in primates (Campbell et al., 2018; Silk, 1994) and rodents (Arakawa, 2018), and these behaviors facilitate social cohesion and bonding. In rodents, barbering (i.e., highly skilled oral plucking of a conspecific’s whiskers/vibrissae) may be used to assert dominance between conspecifics (Sarna et al., 2000). Thus, a variety of touch behaviors are used by social animals to facilitate social cohesion, signal relationship (e.g., kinship) and group ranks.

Human Capacity Data

Several studies—primarily relying on self-report data—suggest that touch frequency (amount used), location (bodily area touched) and quality (type of touch behavior used), may provide an indicator of relationship closeness and power status in humans (Hall et al., 2005; Henley, 1977; Sorokowska et al., 2021; Suvilehto et al., 2015).

  1. (1)

    Touch frequency. Touch frequency may signal relationship phase, closeness and dominance status. In terms of relationship phase, interpersonal touch frequency is highest in the first stage of dating (Emmers & Dindia, 1995; Guerrero and Andersen, 1991), but this effect may be mediated by gender (Willis & Briggs, 1992). Survey evidence indicates that, across cultures, the amount of touch used can signal the closeness of a relationship (Sorokowska et al., 2021). In line with this, greater relationship satisfaction was associated with an increase in self-reported allogrooming in heterosexual relationships, and allogrooming was reportedly more frequent in romantic than close non-romantic relationships (i.e., best-friendships; Nelson & Geher, 2007).

Touch frequency is expected to signal power dynamics within a relationship (e.g., Henley, 1973). Carney et al. (2005) asked participants to imagine two individuals who differed in organizational rank dominance (i.e., boss vs. subordinate), and how much they would expect each person to use 70 different non-verbal behaviors. According to their self-report data, participants expected the individual with the higher dominance rank to instigate more touch behaviors (e.g., hand-shaking, touching) than the subordinate. Thus, touch frequency may be a perceived marker of dominance status in certain contexts.

  1. (2)

    Bodily region touched. Across six cultures, survey data indicated that emotional bond with the toucher explained 54% of the variation in spatial touching patterns (Suvilehto et al.,  2015, 2019). A partner was allowed to touch all body regions, closest acquaintances and relatives were allowed to touch the head and upper torso, and strangers were restricted to the hands (Suvilehto et al.,  2015, 2019).

  2. (3)

    Touch quality. The type of touch used, can also provide an important signal of relationship closeness and power. In terms of relationship closeness, self-report data suggests that affective touch to male and female friends primarily constitutes of hugging and embracing, whereas all forms of affective touch (embracing, hugging, kissing and stroking) are likely used towards children and partners (Sorokowska et al., 2021). Similarly, Hall (1996) examined academics interacting at professional meetings and conventions, and found that while touch frequency was comparable between the higher-status person and lower-status person in a dyad, the higher-status person was more likely to initiate less formal affectionate touches (to the arm or shoulder), whereas the lower-status person was significantly more likely to initiate formal touches (i.e., handshakes).

Thus, these studies provide some indication that the tactile sense plays a powerful role in non-verbal communication of closeness and status, however as the majority are based on self-report data – more field and behavioral studies are warranted to validate these effects.

Human Impaired Data

No human impaired data exists here.

Reproductive Related Functions

Fitness Detection

Animal Data

Little research has examined the role of the tactile sense in gauging the genetic fitness of potential mates. In western lowland gorillas (Gorilla gorilla gorilla) females were more likely to leave the group and migrate to groups with younger males, when the dominant adult male (silverback) had visual skin lesions (Baudouin et al., 2019). This suggests that the visual (colour) and/or tactile (skin texture) cues of the lesions facilitated avoidance of males who were not genetically fit.

Human Capacity Data

There is some evidence to suggest touch may be used for fitness detection of mates. First, soft and smooth feeling skin appears to be a highly desirable trait (Tsankova & Kappas, 2016), with global skin care sales amounting to 155.8 Billion (US) in 2021 (Petruzzi, 2022). Second, humans engage in several tactile based self-grooming behaviors (e.g., picking pimples/skin deformities, shaving and hair waxing) – and this may be done to increase their attractiveness to a potential suitor (McGlone et al., 2016). In line with this, Daly et al. (1983) observed self-grooming behaviors (defined as straightening clothes, hair grooming, gazing at self in a mirror), in individuals at restaurant restrooms, and subsequently interviewed these individuals on their relationship status. They found females self-groomed for longer than males, and individuals who were in a long-term relationship (e.g., married) spent less time grooming than those in the early phase of dating. These findings suggest some tactile as well as visual grooming behaviors may be used to increase perceived attractiveness for mating. Nevertheless, more studies are needed to examine the full range of tactile grooming behaviors used by humans, and the motivations behind their use (McGlone et al., 2016).

Human Impaired Data

No human impaired data exists here.

Intimacy and Sexual Behavior

Animal Data

Prior to sexual activity, in several non-human primates and mammals, the amount of grooming between males and females increases during mating seasons (Burley, 1980; D’Amato et al., 1982; Drickamer, 1976; Matsumoto-Oda, 1999). Female rats increase their foot stomping pre copulation (Burley, 1980), suggesting receptiveness to courtship can be indicated via tactile behaviors in rodents.

Physically mounting the female, during copulation, occurs across several non-human primate species (Michael et al., 1978), as well as other mammals (e.g., rodents; Burley, 1980; rabbits, Contreras & Beyer, 1979). Post-copulatory grooming is commonly observed between non-human primates (Sonnweber et al., 2015; Terry, 1970), and may serve to reduce tension and arousal following copulation or, be used strategically for sex specific reproductive goals (Sonnweber et al., 2015).

Human Capacity Data

Prior to sexual activity, slow touch is perceived as significantly more erotic than fast touch (Bendas et al., 2017). According to self-report data, tactile behaviors are used more commonly than verbal behaviors to express sexual intentions (e.g., Curtis et al., 2012; Vannier & O’Sullivan, 2011). Heterosexual men in relationships self-reported that they were more likely to use kissing and hugging to initiate sexual behavior, whereas heterosexual women in relationships reported they were more likely to use genital touching and removal of clothing (Vannier & O’Sullivan, 2011).

Sexual intercourse and penetration are undeniably reliant on physical contact (Masters & Johnson, 1966). A variety of sexual touches have been identified from survey data (e.g., Willis & Rinck, 1983), indicating certain touch pressures, patterns and styles may be predictive of pleasure in sexual intercourse (Herbenick et al., 2018). Greater tactile sensitivity is related to increased sexual (penile vaginal) intercourse (Brody et al., 2008), as well as increased orgasms following penile vaginal intercourse and enhanced sexual arousal in females (Costa et al., 2017). Greater vibrotactile sensitivity is also related to better erectile function in men – suggesting that tactile sensitivity, across sexes, may be important to sexual arousal and function (Rowland, 1998). Similarly, greater perceived pleasantness of stroking that activates C-LTMRs (affective touch), compared to fast (non C-LTMR related) stroking, was associated with greater sexual desire in women and longer sexual duration in men (Bendas et al., 2017).

Outside of sexual intercourse, several tactile behaviors (backrubs, massages, caressing, stroking, cuddling, holding hands, hugging, kissing on lips and face) relate to participants’ satisfaction in their romantic relationships (Gulledge et al., 2003).

Human Impaired Data

Women with vulvar vestibulitis syndrome (i.e., pain around the vulvar skin), have heightened baseline tactile sensitivity in the vulvar skin region, which is associated with greater pain during sexual intercourse (Payne et al., 2007; Pukall et al., 2000). Notably, both women with, and without, vulvar vestibulitis have greater vulvar tactile sensitivity after viewing an erotic film (Brody et al., 2010). These findings suggest that tactile sensitivity is positively associated with higher sexual arousal in individuals with normative tactile sensitivity, but when tactile sensitivity is too high at baseline – especially in the vulvar region – this may characterize sexual pain disorders and dysfunction during intercourse.

Caregiver and Infant Attachment

The tactile sense is the first sense to develop in the human infant. At four weeks old gestation, cutaneous and trigeminal functions begin to develop in the human infant, with haptic skills (i.e., grasping and rooting reflexes) developing by 12 weeks (Borsani et al., 2019; Bremner & Spence, 2017). By contrast, visual and auditory functions do not begin to develop until 24 weeks gestation. Thus, relative to the other sensory modalities, the newborn infant is most responsive to tactile stimuli in the environment (Bremner & Spence, 2017; Hertenstein, 2002). A large body of scientific work has dedicated itself to understanding the role of caregiver touch and touch deprivation in social communication and emotional development (see Bremner & Spence, 2017; Cascio et al., 2019; Gallace & Spence, 2010; Hertenstein, 2002; Morrison & Croy, 2021 for reviews). Below, we turn to prior reviews and subsequent research on the role of the tactile sense, across species, in caregiver and infant attachment.

Animal Data

The tactile sense plays a critical role in the formation of infant and caregiver attachment and subsequent social development in non-human primates (Griffin & Harlow, 1966; Harlow, 1958; Harlow & Zimmerman, 1959; Harlow et al., 1965). Harlow and colleagues' (1958; 1959; 1965) work on non-human primates, found that tactile qualities (i.e., warmth, softness) of the caregiver, rather than their ability to afford food or protection were the most important features for ensuring caregiver and infant attachment (aka contact comfort). Since Harlow’s work, researchers have shown that non-human primate infants display increased stress symptoms, following brief separation (i.e., 30 min to 10 days) from their caregiver (Coe, et al., 1978; Reite et al., 1981). The tactile sense is pertinent for facilitating caregiver attachment in rodents (e.g., Caldji et al., 1998, 2000; Kaffman & Meaney, 2007; Liu et al., 1997), birds (e.g., Clements & Lien, 1975; Eiserer, 1978; Gottlieb, 1993), elephants (e.g., Lee, 1987) and ungulates (e.g., Nowak and Boivin, 2015). Thus, the role of contact comfort in establishing filial bonds, social and affective development, seems conserved across several species.

Human Capacity Data

Body proximity and contact are central to the formation of a secure attachment style (Ainsworth, 1979; Bowlby, 1982; Bretherton, 1992), in which the infant seeks close proximity and contact with the mother, post separation. By contrast, infants with anxious-avoidant or anxious-ambivalent attachment styles, respectively display avoidance of the mother, or unpredictable contact following reunion with their caregiver (Ainsworth, 1979; Ainsworth & Blehar, 1975; Anisfeld et al., 1990; Main & Cassidy, 1988).

The quality and frequency of physical contact predicts the security of the caregiver-infant attachment and in turn, the infants’ social-affective development (Bell & Ainsworth, 1972). Ainsworth and Blehar’s (1975) observation of 26 mothers with their infants revealed that tender, careful holding of the infants was most predictive of positive responses in the infant. Similarly, Anisfeld et al. (1990) demonstrated that infants who were carried by their mothers using soft baby carriers (more physical contact) were significantly more likely to be securely attached at 13 months, than infants who were carried using infant seats (less physical contact).

Caregiver touch appears pertinent to facilitating positive affect and reducing negative affect in the infant. Tanaka et al. (2021) found that infants placed in a high physical touch group with their mothers, had significantly greater displays of positive affect, less avoidance of strangers and greater object exploration, than infants placed in a low physical touch condition. Maternal touch can reduce crying and cortisol levels in infants, and increase smiles and eye contact with the mother, even when the mother is staring expressionless at their infant (Feldman et al., 2010; Stack & Muir, 1990, 1992). Thus, caretaker touch may reduce negative affect induced from other non-verbal behaviors.

Tactile cues may be used in kin recognition. Mothers can discriminate between the feel of their infant’s skin (i.e., with only 1–29 h of infant contact) from a stranger’s infant (Kaitz et al., 1993). Similar results have been replicated with fathers (Kaitz et al., 1994), who reported using tactile cues, such as smoothness, fattiness or boniness of hand, and hand size, to determine which infant was their own (Bader and Philips, 1999).

Human Impaired Data

Spitz’s (1945) observations on the factors responsible for unfavorable infantile development in institutions demonstrated the debilitating effects of maternal touch deprivation in human infants. More recently, social touch researchers have examined if aversion to caretaker touch is predictive of autism spectrum disorder. Baranek (1999) examined equal duration home videos of typically developing children and children diagnosed with autism spectrum disorder, at the first 9–12 months of the child’s life (i.e., prior to autism spectrum diagnosis). Children with autism spectrum disorder had significantly greater aversion to social touch, compared to typically developing children (and see Mammen et al., 2015). Thus, retrospective data indicates that the relation between caregiver and infant touch may be an important predictor of a child’s later social and cognitive development.

Overall Discussion

Across human (capacity, impaired) and animal studies, the tactile sense appears to have a conserved role in all three of its functional domains: (1) Ingestive Behavior; (2) Environmental Hazard Detection and Management, and (3) Social Communication (see Fig. 1, for a summary). While this is the first attempt to categorize all the behavioural functions of touch, the identified categories appear pertinent to other modalities too (see Stevenson, 2010a, for olfaction). The significance of an integrative review in understanding the unique value of touch compared to the other senses, its versatility in non-verbal communication and social psychology more generally, are turned to below.

In the Introduction, we argued that an appraisal of function allows us to determine the unique contribution of the tactile sense (relative to other sensory modalities), for each function broadly construed (e.g., disease avoidance in general). However, addressing this is complicated for touch. First, the tactile sense is not localized to one body area unlike other modalities – i.e., sight is localized to the eyes, olfaction to mouth/nose, taste to mouth, and hearing to ears. By contrast, tactile sensation not only occurs on the skin, but also in tandem with other senses – e.g., texture can be perceived visually (bubbly surface of crispy chips), orally (crispy texture in mouth) and auditorily (hearing crunchiness). Complete loss of tactile perception may thus involve multimodal sensory loss, making it highly unlikely (Spence, 2013; Stevenson, 2010b; Sun et al., 2016). In line with this, there appears to be no reported case of global tactile loss – however, rarely, few classes of touch fibres – discriminative (Olausson et al., 2002), C-LTMR (Morrison et al., 2011), pain (Nagasako et al., 2003) – can be lost or severely reduced. This suggests, perhaps, a more feasible direction to study functional contribution, namely, to examine how the tactile sense may compensate for loss in the other modalities (e.g., Alary et al., 2008).

Relatedly, the functions identified in this review parallel those identified for the human olfaction system (Stevenson, 2010a), potentially because these functions appear pertinent to self (ingestion, hazard detection) and species (social communication) preservation. It is interesting then to speculate if these functions could be identified for the other modalities. The role of vision in ingestive behavior (e.g., Spence, 2013; Stevenson, 2010b), hazard detection (e.g., Croy et al., 2013; Regenbogen et al., 2017; Sarolidou et al., 2020) and social communication (e.g., Darwin, 1872; Ekman, 1993) is suggested in several places. Taste is clearly of importance to ingestive behavior (Spence, 2013) and hazard detection (Chapman & Anderson, 2012) – however, its role in social communication is less apparent and more metaphorical (e.g., ‘he’s sweet’). While audition has a role in social communication (Jackson & Turbull, 2004), it may have a more limited role in ingestive behavior (Oleszkiewicz et al., 2023; Spence & Shankar, 2010) and disease related hazard detection (Croy et al., 2013; Dominy et al., 2001; Lamond et al., 2024). A more systematic appraisal of whether these functions can be applied across senses, would both highlight the unique and relative cost of injury to each sensory system, as well as determining if these three encompass all core functions.

One benefit of an integrative review of tactile function, compared to a review of a single function, is that it can provide insight on the versatility of the tactile system and where non-verbal behaviors may have evolved from. In line with this, there are several parallels between different functions and sub-functions reviewed (as illustrated in Fig. 1). For instance, the role of the tactile sense in nipple localization during breast-feeding, appears not only pertinent to ingestive behavior but also to infant caregiver attachment – i.e., the infant learns to use touch first to communicate their physical needs, and then their emotional needs to the caregiver (Bremner & Spence, 2017; see Fig. 1). Another, is the overlap between hazard detection and affective communication behaviors. To protect ourselves from disease related threats, humans use grooming, scratching and wiping behaviors, which appear to be motivated, at least in part, by the emotion disgust (Blake et al., 2017; Kupfer & Fessler, 2018; Oum et al., 2011; Saluja & Stevenson, 2019). Similarly, when individuals are asked to communicate disgust to other individuals (in interpersonal communication) pushing and wiping behaviors are also adopted (Hertenstein et al., 2006a; see Fig. 1), indicating that this disease avoidance drive may form the basis for some forms of affective communication. In sum, tactile (non-verbal) behaviors can map to several functions, which may inform about the way in which these behaviors evolved.

A final issue to consider then, is how to bind together the various tactile functions. From a social psychological perspective, a common theme appears to be affect. Unwanted intimate touch, eating a cold soggy pie, and the feel of treading in warm dog faeces, all generate sensations that evoke potent negative affects. On the other hand, welcome sexually intimate contact, gourmet food, and a warm bath can all evoke the opposite. Affectivity is thus pertinent to touch, and the strong positive and negative emotions, drives and desires that accompany real or imagined tactile sensations. Several testable implications flow from this observation. First, it should be possible to deliberately deploy the tactile sense to manipulate positive or negative affect in social related behaviors (e.g., Williams & Bargh, 2008; but see Lynott et al., 2014). Second, utilizing tactile behaviors to this end may operate below the level of conscious awareness, because many of these effects would probably never have been explicitly noticed. Third, there may be interesting commonalities across function, such as food avoidance, dislike of touch and overactive disgust – as observed in anorexia nervosa – suggesting a general dysfunction in the tactile sense. Finally, it is noteworthy, that in the examples used at the start of this section, the interpretive frame can often be of great significance in dictating whether essentially the same tactile sensory experience evokes positive or negative affect (e.g., contrast an intimate caress from a stranger with that from a lover). This ability for interpretive frame to dictate the affective response is only starting to be examined (e.g., Gruhl et al., 2022; Saluja and Stevenson, 2022), but offers a new experimental window for studying tactile illusions, and how the same stimulus can evoke such different responses.

In conclusion, there is good evidence to suggest that Ingestive Behavior, Environmental Hazard detection and Social Communication are conserved functions of the tactile sense across species, and for humans, and that they share a unifying theme, in their capacity to powerfully evoke affect.