The self-perceived body size in obesity: Evidence from the implicit representation of the hand

Abstract

To locate our body in the space, we rely on an implicit representation of body size and shape: the body model. Evidence about the implicit representation of bodily dimensions in obesity is rare. Nevertheless, it seems to suggest that such representation is not altered in obesity compared to healthy weight individuals. To probe further this hypothesis, we investigated the implicit representation of hand dimensions with a landmark localisation task, comparing individuals with obesity and healthy weight individuals. Furthermore, as body model distortions may be related to tactile acuity, the tactile acuity threshold was measured using a two-point discrimination task. In accordance with the previous evidence, we observed that healthy weight participants showed a significant underestimation of finger length and overestimation of hand width. Interestingly, comparable body model distortions were shown also in participants with obesity. No differences in tactile acuity emerged between the two groups; also, when considering the whole sample, as tactile acuity decreases hand width overestimation increases. Thus, obesity seems to have no effect on the characteristics of the body model relative to the hand. Accordingly, the physiological mechanisms supporting the development of the implicit representation of hand dimensions in the healthy weight population may be preserved in obesity.

Keywords

Obesity hand size estimation body model implicit representation position sense tactile acuity

Introduction

Suppose we want to locate the position of our own hand in the external space without seeing it. Peripheral somatosensory receptors inform our brain about the degree of flexion of the shoulder, the elbow, and the wrist (Proske & Gandevia, 2012), outlining the reciprocal arrangement of the arm, the forearm, and the hand (Burgess et al., 1982). Hence, the position of the hand relative to the other connected body parts is defined. However, to figure out its absolute position, namely where the hand is exactly located, the length of these body segments must be known (Longo et al., 2010; van Beers et al., 1998). Thus, to properly localise the body in the external space, the somatosensory information must be combined with a pre-stored knowledge about the body metric. This implicit (i.e., inaccessible to conscious awareness) representation of bodily dimensions has been called body model (Longo & Haggard, 2010).

In 2010, Longo and Haggard designed a landmark localisation task to investigate the body model of the hand. In this task, the participant puts his or her own hand under a horizontal panel and localise the position of each fingertip and knuckle using a tool, such as a baton. The perceived position of fingertips and knuckles is used to compute a structural representation of the metric properties of the hand. Specifically, the distance between the perceived location of each fingertip and knuckle is used to estimate the internal representation of the length of each finger, while the distance between the perceived location of the index knuckle and the little finger knuckle is used to estimate the internal representation of the hand width.

So far, several studies reported that healthy individuals systematically underestimate the finger length, incrementally from the thumb to the little finger, and overestimate the hand width in the landmark localisation task (Coelho & Gonzalez, 2018; Longo & Haggard, 2010; Mattioni & Longo, 2014; Saulton et al., 2016; Tamè et al., 2017). Thus, the body model seems to be slightly inaccurate.

Altered perceptions of body size have been reported in obesity (e.g., Schwartz & Brownell, 2004), a medical condition characterised by an excessive accumulation of body fat (World Health Organization, 2020). These studies largely focused on the conscious, and thus explicit, perceptual, conceptual, and emotional representation of the body (Schwartz & Brownell, 2004), that is the body image (Gallagher, 2005; Head & Holmes, 1911; Paillard, 1999). For example, participants affected by obesity were asked to draw lines representing the horizontal width of some body parts (e.g., Scarpina et al., 2017), to place a piece of rope in a circle to estimate their circumference (e.g., Scarpina et al., 2019), or to compare the dimensions of their own body with a picture (e.g., Johnstone et al., 2008) or a silhouette (e.g., Bell et al., 1986). These tasks purport to measure how individuals think their own body dimensionally: the comparison between the reported body size and the actual body size represents the magnitude and the direction of the possible distortions of body representation.

Nevertheless, the concept of body image is different from the concept of body model: even though they might share some characteristics (in principle, both are representations of the size and shape of the body), the “body model shows large distortions which critically do not appear to characterise the body image” (Longo, 2015c). Furthermore, we might observe that the emotional components which are critically included in body image are not taken into account in the concept of body model. Indeed, any emotion or belief related to one’s own body is more likely to affect body size estimation while individuals are aware about judging their body (i.e., body image) than when they do it unconsciously by recruiting an implicit body representation (i.e., body model).

Previous evidence on the implicit representation of bodily dimensions in obesity is quite scarce and none of the investigations referred to the concept of body model (Longo & Haggard, 2010). Scarpina and colleagues (2014) investigated the accuracy of participants affected by obesity and healthy weight individuals in estimating the distance between tactile stimuli applied on the forearm and the abdomen. The authors reported that both groups implicitly overestimated tactile distances, even though in obesity this overestimation was significantly higher when the stimulation was delivered on the forearm than on the abdomen (Scarpina et al., 2014). Also, both children and adolescents with obesity and with healthy weight overestimated tactile distances applied on the arm, the spine, the thigh, and the buttocks, but the estimation bias was significantly higher in affected individuals for all body parts, except the arm (Mölbert et al., 2016). The estimation of the distance between tactile stimuli requires to refer implicitly to a pre-stored representation of the bodily dimensions (Longo et al., 2010; Longo & Haggard, 2011; Spitoni et al., 2010), likewise for position sense. In brief, if one (implicitly) knows that a certain body segment is long n, and the stimulus is covering half of that surface, the size of the stimulus will be estimated as n/2. As a result, any distortions in the estimation of the distance between tactile stimuli can be interpreted as a distortion in the underlying implicit representation (Longo et al., 2010; Longo & Haggard, 2011). Accordingly, the previous results are suggestive of a distorted implicit representation in both obesity and healthy weight, even though in obesity this bias might be increased for specific body parts. Nevertheless, it might be considered that tactile distances are generally perceived longer when oriented along the medio-lateral (i.e., horizontal) than the proximo-distal (i.e., longitudinal) body axis (e.g., Longo et al., 2015; Longo & Haggard, 2011; Longo & Morcom, 2016; Longo & Sadibolova, 2013; Tamè et al., 2017), but these anisotropies were not reported for the abdomen (Green, 1982; Longo et al., 2019; Marks et al., 1982). In the mentioned studies, tactile distances have been delivered vertically on the arm and horizontally on the abdomen. Thus, different results might be expected if considering horizontal tactile distances on the forearm (but no differences are expected if considering vertical distances on the abdomen).

Lately, Scarpina and colleagues (2017) focused on the implicit representation of body metric involved in action. When passing through a door, our movements must be calibrated according to both the dimension of the body and the (visual) size of the aperture (Keizer et al., 2013) to avoid collisions. Hence, the perception of affordances (i.e., of passing or not through a door of a certain aperture) is tightly related to the (implicitly) perceived body size, so much so that it can be quickly modulated after temporary identification with a virtual larger or thinner body avatar (Piryankova et al., 2014). Accordingly, the shoulders (Keizer et al., 2013) and/or hip (Scarpina et al., 2017) range of motion may represent a clue about possible distortions of body representation. Individuals with obesity, likewise healthy weight participants, moved according to their physical dimensions when they passed through different apertures, suggesting an accurate underlying (implicit) representation of the bodily dimensions (Scarpina et al., 2017).

Therefore, on one side, the previous findings are suggestive of overall similar implicit representations of bodily dimensions in obesity and healthy weight; on the other side, differences (i.e., higher distortions) can be detected in obesity when considering certain body parts, at least in a tactile distance estimation task. However, none of the mentioned studies investigated the body model underlying position sense. Thus, to further test the hypothesis of a similar implicit representation of bodily dimensions in obesity and in healthy weight, we used an adapted version of the landmark localisation task (Longo & Haggard, 2010). According to the previous literature, we may expect healthy weight individuals to underestimate the finger length and overestimate the hand width (Coelho & Gonzalez, 2018; Longo & Haggard, 2010; Mattioni & Longo, 2014; Saulton et al., 2016; Tamè et al., 2017). Furthermore, we expect individuals with obesity and healthy weight individuals to have a qualitatively similar (distorted) body model, although the magnitude of the estimation bias might be higher in obesity.

Yet, one might ask why the body model is distorted. Longo and Haggard (2010) related the estimation biases of the hand dorsum in the localisation task to the fact that tactile acuity is higher on the medio-lateral axis (which is overestimated) than on the proximo-distal axis (which is underestimated). In other words, they suggested that the body model distortions might be related to anisotropies in tactile acuity. Later studies showed that the dimensions of body parts with higher tactile acuity are perceived more accurately than those with lower tactile acuity when participants explicitly estimate the body size relative to either other body parts or objects (Linkenauger et al., 2015; Sadibolova et al., 2019). Also, Peviani and colleagues (2019) found that, when participants estimate the dimensions of the hand relative to visually presented lines, body parts with lower tactile acuity were perceived longer (i.e., underestimated less) than body parts with higher tactile acuity. Furthermore, temporary alterations of tactile sensitivity induced by local anaesthesia determined body size overestimation (Proske & Gandevia, 2012). Thus, estimation biases might be related also to tactile acuity per se.

We might note that the mentioned studies used more explicit tasks than the landmark localisation task; however, the task used in Peviani and colleagues (2019) is supposed to involve, at least partially, an implicit representation of the body metric whose distortions are similar but less pronounced than in the body model underlying position sense (Longo & Haggard, 2012b). Indeed, the distinction between implicit and explicit body representations might not be categorical, with body representations lying on the same continuum and possibly interplaying according to the task demands (Longo, 2015a; Longo & Haggard, 2012b).

Therefore, one might cautiously speculate that tactile acuity per se may affect also the body model distortions. This might be specifically relevant in the case of our study, considering the lower tactile acuity reported in obesity than in healthy weight participants (Boles & Givens, 2011; Bussolaro et al., 2012; Falling & Mani, 2016). Thus, a secondary aim of our work was to probe the relationship between tactile acuity and the body model distortions, measuring the tactile acuity threshold with a two-point discrimination task in both groups (Falling & Mani, 2016; Klein, 2001).

Methods

Participants

The study was approved by the ethical committee of the I.R.C.C.S. Istituto Auxologico Italiano and was performed in compliance with the ethical standards laid down in the Declaration of Helsinki (recently amended in Fortaleza, 2013). All participants were naïve to the rationale of the experiment and gave their informed written consent. Participants affected by obesity were inpatients at the I.R.C.C.S. Istituto Auxologico Italiano, Ospedale San Giuseppe (Piancavallo, Oggebbio, VCO, Italy), recruited before attending a weight-loss rehabilitative programme. Healthy weight individuals were recruited from the University of Trento or by experimenters’ contact; they participated in return for academic credits or monetary compensation (7 euros).

A convenience sample was recruited due to the time constraints imposed by the clinical context; thus, we did not perform a priori power analysis to define sample size. To compensate for this limitation and quantify the uncertainty of our results, we reported 95% standard symmetric confidence intervals (CIs) around the observed effect size for parametric and non-parametric tests (see Thomas, 1997). Fifteen participants affected by obesity with a body mass index (BMI, kg/m²) equal or higher than 30, 9 females; age in years: M (SD) = 46 (8); education in years: M (SD) = 11 (4); BMI in kg/m²: M (SD) = 42.77 (5.21), and 19 healthy weight individuals with a BMI lower than 30, 16 females; age in years: M (SD) = 44 (9); education in years: M (SD) = 14 (3); BMI in kg/m²: M (SD) = 23.98 (4.21), were enrolled. All participants were right-handed, as they self-reported. The presence of any neurological, motor, and/or sensory impairment was an exclusion criterion.

Independent samples t-tests revealed that the two groups had comparable age, t₍₃₂₎ = 0.75, two-tailed p = .46, d = 0.26, 95% CI = [–0.42, 0.94,], while healthy weight participants had a higher level of education, t₍₃₂₎ = 2.52, two-tailed p = .02, d = 0.87, 95% CI = [0.16, 1.57], in line with previous studies (Scarpina et al., 2014, 2017). As expected, individuals with obesity had a higher BMI than healthy weight participants, t₍₃₂₎ = –11.64, two-tailed p = .001, d = 4.02, 95% CI = [2.82, 5.20].

Tasks and procedure

Two tasks were performed in counterbalanced order across participants, such that half of participants in both groups started with the landmark localisation task, while the other half began with the two-point discrimination task.

Landmark localisation task

An adapted version of Longo and Haggard’s (2010) task was administered. The participant was blindfolded and sat comfortably in front of a table, on which a 23-inch LCD touch-screen monitor (HP 2310ti) was placed horizontally, nestled in a polystyrene support. The polystyrene apparatus was held up by four wooden supports with the screen standing 12 cm over the table. The participant put the right hand under the screen on a velvet patch, palm down with the fingers spread in a comfortable position. The hand was located approximately 30 cm from the body, aligned with the body midline. The left hand was placed on the polystyrene apparatus on another velvet patch 10 cm to the left of the screen border, aligned with the shoulder. The velvet patch on the apparatus indicated the point where the participant had to return the hand after each trial (i.e., the starting position) (Figure 1a).

Figure 1.

(a) Illustration of the set-up and the procedure of the landmark localisation task. (b) Example of the spatial map of the hand.

In each trial, the participant had to locate the fingertip or the knuckle of one of the fingers of the right hand touching the screen with the index of the left hand. More specifically, the participant was instructed to touch the screen in correspondence of the spatial position in which she or he proprioceptively perceived the landmark. Ten landmarks were probed: the fingertip and the knuckle (i.e., metacarpophalangeal joints) of the thumb, the index, the middle, the ring, and the little finger of the right hand. Before the task, the experimenter touched each landmark to ensure a proper understanding of the targets to be located. The indication of which landmark to pinpoint was given through headphones before each trial (e.g., “index fingertip”; “middle finger knuckle”). When the participant touched the screen, a sound signalled the recording of the location by the system, while not providing any feedback about the accuracy of the localisation. Trials delivery was self-paced: the participant was encouraged to take her or his time, trying to be as accurate as possible. The participant was also told to judge each landmark independently from the previous ones, avoiding any strategy to relate consecutive judgements. To favour so, the participant had to return the left hand to the starting position after each trial. Each landmark was localised 10 times: overall, the experiment comprised 100 trials. Trials order was pseudo-randomised within participants, with no consecutive repetition of the same target landmark.

The x-y pixel coordinates of each point were recorded using E-Prime 1.2 software (Psychology Software Tools, Pittsburgh, PA). For each participant, the mean of the x-y pixel coordinates relative to the same landmark, across the 10 repetitions, was computed. Following Longo and Haggard’s (2010) procedure, a spatial map of the reciprocal locations of each fingertip and knuckle was obtained (Figure 1b). Then, the distances between the averaged pointed landmarks were converted from pixels to centimetres. The distance between each averaged pointed fingertip and the corresponding averaged pointed knuckle was used as an implicit measure of the represented finger length. Instead, the distance between the averaged pointed knuckles of the index and the little finger was used as an implicit measure of the represented hand width. Thus, we obtained an indirect measure of the metric of the implicit body model of the hand. At the end of the experiment, the real dimensions of the participant’s hand were measured with a ruler.

Two-point discrimination task

Pairs of two-point tactile stimuli were delivered on the centre of the right-hand dorsum, perpendicularly to the proximo-distal body axis, approximately 1 cm below the lowest knuckle of the middle finger. The participant had to verbally report whether she or he perceived one point, two points or whether she or he was unsure about the perception (i.e., three-alternative forced-choice method; see, for example, Klein, 2001). Stimuli were delivered using the tips of a commercially available digital calliper (Falling & Mani, 2016). The inter-stimulus interval was nearly 5 s, while the stimulation lasted nearly 2 s.

Stimuli were delivered according to the psychophysical adaptive staircase method with an up-down tracking rule (Cornsweet, 1962; Ehrenstein & Ehrenstein, 1999). For all participants, the starting distance between the two tips of the calliper was set at 30 mm (Mancini et al., 2014). According to Falling and Mani (2016), the distance between the tips decreased of one step in each successive trial, as long as the participant perceived two points. Conversely, when the participant perceived only one stimulus, or she or he was unsure, the distance between the tips increased of two steps. Each transition between the detection of two tactile stimuli and the detection of one stimulus represented a “reversal”: this point indicates the minimum distance required to perceived two stimuli separately, namely the tactile acuity threshold (Cornsweet, 1962; Ehrenstein & Ehrenstein, 1999). The procedure lasted until 12 reversals were recorded. A step of 5 mm was used in the first six reversals; a step of 2 mm was adopted in the last six reversals (Falling & Mani, 2016).

Results

Landmark localisation task

Percentage estimation errors were computed according to the following formula:

[(estimated size_real size) / real size)] \times 100

Negative values indicate an underestimation; positive values an overestimation. An estimation error equal to zero means that the estimated dimension corresponds to the real one. Estimation errors were computed for the judged length of each single finger and the hand width. Single fingers estimation errors were averaged into one index, which represents the overall finger length estimation error (i.e., the average finger length).

Finger length

On average, finger length was underestimated in both groups, healthy weight: M (SD) = –21.39% (21.57); obesity: M (SD) = –17.12% (26.50). Within-groups one-sample t-tests on normal data (Shapiro–Wilk test: healthy weight p = .87; obesity p = .09) revealed that the average finger length was significantly different from zero in both healthy weight, t₍₁₈₎ = –4.32, two-tailed p = .001, d = 0.99, 95% CI = [0.43, 1.54], and the obesity group, t₍₁₄₎ = –2.50, two-tailed p = .03, d = 0.65, 95% CI = [0.08, 1.20], as shown in Figure 2 (left panel).

Figure 2.

Left panel: mean (bars) and standard error (vertical lines) of the percentage estimation error (y-axis) of finger length (on the left) and hand width (on the right) in healthy weight group (dark grey) and in the group with obesity (light grey). Positive values indicate an overestimation; negative values, an underestimation. Asterisk indicates the magnitude of the error was significantly different from zero (i.e., accuracy level). Right panel: the mean (bars) and standard error (vertical lines) of the tactile acuity threshold in mm (y-axis) in healthy weight (dark grey) and in the group with obesity (light grey).

Furthermore, the finger length underestimation was comparable between the two groups, as shown by an independent samples t-tests, t₍₃₂₎ = –0.52, two-tailed p = .61, d = 0.18, 95% CI = [–0.50, 0.86], suggesting no effect of the group on the average finger length estimation. To probe further this hypothesis (i.e., the null hypothesis), a Bayesian two-sided independent samples t-test was performed using JASP (JASP Team, 2019; JASP, Version 0.11.1) to compare the average finger length between obesity and healthy weight. In fact, Bayesian analyses can discriminate between the “absence of evidence” (e.g., not enough data) and the “evidence of absence” of any effect, namely the evidence in favour of the null hypothesis (Dienes, 2014). Given that this was the first study comparing healthy weight and individuals with obesity in the landmark localisation task, no previous data were available to estimate the expected effect size; therefore, the value settled by the software for the prior was used. The prior was described by a Cauchy distribution centred around zero with a width parameter of 0.707. This corresponds to a probability of 80% that the effect size (computed as μ/σ, where μ represents the population mean and σ its standard deviation; Ly et al., 2016; Rouder et al., 2009) is between −2 and 2.The analysis revealed a weak/nearly moderate support for the null hypothesis (b₀₁ = 2.73; b₁₀ = 0.37; Jeffreys, 1961), thus, consolidating the previous result about no effect of the group on the finger length underestimation.

Non-parametric Spearman’s coefficients of correlation were computed between the BMI and the average finger length to probe the possible relationship between body mass and finger length underestimation, in each group independently. The analysis showed that participants’ BMI did not significantly correlate with the average finger length neither in obesity (r = .18, p = .52, 95% CI = [–0.36, 0.64]) nor in the healthy weight (r = –.05, p = .85, 95% CI = [–0.49, 0.42]; see Figure 4a).

Finally, the possible radio-ulnar gradient (i.e., increasing underestimation from the thumb to the little finger; see, for example, Longo & Haggard, 2010) was analysed. According to previous literature (Ganea & Longo, 2017; Longo, 2014, 2015b, 2017, 2018; Longo et al., 2012; Longo & Haggard, 2010, 2012a, 2012b; Mattioni & Longo, 2014; Tamè et al., 2017), for each participant, a least-squares regression was performed: the underestimation of each finger was set as dependent variable; fingers from the thumb to the little finger represented the predictors (a progressive number from one to five was assigned to each finger to run the analysis). Thus, the regression coefficient (b) represents the increment (if negative) or decrement (if positive) in underestimation from one finger to the consecutive one, in the radial-ulnar (i.e., thumb–little finger) direction. A regression coefficient equal to zero indicates no increment in underestimation across the fingers. To compare the two groups, the average regression coefficients were computed as the mean of the regression coefficients of all participants in each group.

Qualitatively, the magnitude of finger length underestimation increased from the thumb to the little finger in both groups (see Figure 3), obesity: average b (SD) = –1.34% (8.78) per finger; healthy weight: average b (SD) = –2.38% (5.46) per finger. Refer to Table 1 for details. However, this increment was not statistically different from zero in both healthy weight, t₍₁₈₎ = –1.90, two-tailed p = .07, d = 0.44, 95% CI = [–0.04, 0.90], and obesity, median b = –2.11%, z = –1.36, two-tailed exact p = .19, r_b = –0.40, 95% CI = [–0.76, 0.15], according to the one-sample t-test performed on normal data (Shapiro–Wilk test p = .33) in the healthy weight and to the Wilcoxon signed-rank test performed on non-normal data (Shapiro–Wilk test p = .01) in obesity. Furthermore, a Mann–Whitney U test showed that the average regression coefficients were comparable between the two groups (U = 127, z = –0.54, two-tailed exact p = .61, r_b = –0.11, 95% CI = [–0.47, 0.28]).

Figure 3.

Mean (bars) and standard error (vertical lines) of the percentage estimation error (y-axis) of each finger in the healthy weight group (dark grey) and in the group with obesity (light grey).

Table 1.

Mean and standard deviation (in brackets) of single finger estimation errors in healthy weight and obesity.

Finger	Healthy weight	Obesity
Thumb	−15.20 (34.01)	−14.96 (38.92)
Index	−20.36 (21.89)	−16.67 (29.20)
Middle	−21.64 (22.46)	−13.71 (26.52)
Ring	−24.93 (22.20)	−20.55 (30.59)
Little	−24.83 (20.08)	−19.71 (34.76)

Hand width

Hand width was significantly overestimated in both groups: within groups one-sample t-tests on normal data (Shapiro-Wilk test: healthy weight p = .08; obesity p = .84) revealed that hand width overestimation was significantly different from zero in both healthy weight, M (SD) = 66.06% (41.62); t₍₁₈₎ = 6.92, two-tailed p = .001, d = 1.59, 95% CI = [0.89, 2.261], and obesity, M (SD) = 65.96% (42.04); t₍₁₄₎ = 6.08 two-tailed p = .001, d = 1.57, 95% CI = [0.79, 2.32] (see Figure 2, left panel). No significant difference between the two groups in the estimation error of the hand width, t₍₃₂₎ = 0.007, two-tailed p = .96, d = 0.002, 95% CI = [–0.68, 0.68], was observed, suggesting no effect of the group on the hand width estimation. As for the average finger length, a Bayesian two-sided independent samples t-test was performed to verify whether our results are in support of the null hypothesis (i.e., no differences between the two groups) or rather denote a lack of evidence. The same procedure used for the average finger length was applied using JASP (JASP Team, 2019; JASP, Version 0.11.1). The analysis indicated a moderate support for the null hypothesis (b₀₁ = 3.03; b₁₀ = 0.33; Jeffreys, 1961), confirming that the group has no effect on the hand with estimation error.

Finally, non-parametric Spearman’s coefficients of correlation were computed between BMI and hand width estimation error to probe the possible relationship between body mass and the experimental outcome, in each group independently. The analysis revealed that the BMI did not significantly correlate with the hand width estimation neither in obesity (r = .21, p = .44; 95% CI = [–0.34, 0.66]) nor in healthy weight (r = .31, p = .19, 95% CI = [–0.17, 0.67]; see Figure 4b).

Figure 4.

Scatterplots representing the percentage estimation errors of the (a) average finger length, (b) hand width, and the (c) tactile acuity threshold according to BMI in the healthy weight group (dark grey) and in the group with obesity (light grey).

Two-point discrimination task

For each participant, the tactile acuity threshold (i.e., the minimal distance at which two tactile stimuli can be perceived separately; Cornsweet, 1962; Ehrenstein & Ehrenstein, 1999) was computed as the average of the distance between the calliper tips across the last six reversals (Falling & Mani, 2016). The independent samples t-test performed on normally distributed data (Shapiro–Wilk test: healthy weight p = .77; obesity p = .24) revealed that the tactile acuity threshold was not significantly different, t₍₃₂₎ = –0.98, two-tailed p = .33, d = 0.34, 95% CI = [–0.34, 1.02], between healthy weight participants, M (SD) = 14.99 mm (5.07), and participants with obesity, M (SD) = 16.82 mm (5.62) (Figure 2, right panel). Therefore, as explained in the previous analyses for the average finger length and hand width estimation errors, a Bayesian two-sided independent samples t-test was performed to verify whether our results actually support the hypothesis of no differences between the two groups in terms of tactile acuity. The analysis revealed a weak support for the null hypothesis (b₀₁ = 2.07; b₁₀ = 0.48; Jeffreys, 1961).

Finally, non-parametric Spearman’s coefficients of correlation were computed between BMI and the tactile acuity threshold to probe the possible relationship between body mass and tactile acuity, in each group independently. The analysis revealed that the BMI did not significantly correlate with the tactile acuity threshold neither in obesity (r = –.22, p = .43, 95% CI = [–0.66, 0.33]) nor in healthy weight (r = .27, p = .27, 95% CI = [–0.21, 0.64]; see Figure 4c).

Tactile acuity and body model distortions

First, non-parametric Spearman’s coefficients of correlation were computed in each group between the average finger length/hand width estimation errors and the tactile acuity threshold to investigate the possible relationship between somatosensation and the body model distortions. However, the tactile acuity threshold was not significantly correlated with either the finger length or hand width estimation in both groups, obesity (fingers length: r = .28, p = .31, 95% CI = [–0.27,0.69]; hand width: r = .22, p = .43, 95% CI = [–0.33, 0.66]) and healthy weight (fingers length: r = .32, p = .19, 95% CI = [–0.16, 0.67]; hand width: r = .34, p = .16, 95% CI = [–0.14, 0.69]; see Figure 5).

Figure 5.

Scatterplots representing the percentage estimation errors of the average finger length and hand width according to the tactile acuity threshold in the healthy weight group (dark grey) and in the group with obesity (light grey).

Thus, a general linear model (GLM) was fitted to our data with either the average finger length or hand width estimation errors as dependent variable, group (obesity vs healthy weight) as between-subjects independent variable and the tactile acuity threshold as covariate. In this way, the possible main effect of both group and tactile acuity on body size estimations can be considered simultaneously within the whole sample. Homogeneity of variances and normality of residuals distribution were tested and found to be guaranteed for both the average finger length (Levene’s test p = .16; Shapiro–Wilk test: obesity p = .44, healthy weight p = .16) and hand width (Levene’s test p = .35; Shapiro–Wilk test: obesity p = .38, healthy weight p = 1.00). For the average finger length, neither the group, F _(1,31) = 0.06, p = .81, partial η² = .002, nor tactile acuity threshold, F _(1,31) = 2.5, p = .12, partial η² = .08, had a significant main effect. For the hand width, no significant main effect of the group was reported, F_(1,31) = 0.16, p = .70, partial η² = .01, but a significant main effect of the tactile acuity threshold was observed, F_(1,31) = 4.98, p = .03, partial η² = .14, suggesting that the hand width overestimation increases as the tactile acuity threshold increases (i.e., tactile acuity decreases; standardised β = .39), when considering the whole sample.

Discussion

We aimed to investigate the implicit representation of the hand underpinning position sense, the body model, using the landmark localisation task (Longo & Haggard, 2010) in obesity. To the best of our knowledge, no previous study probed this issue. According to the literature (Coelho & Gonzalez, 2018; Longo & Haggard, 2010; Mattioni & Longo, 2014; Saulton et al., 2016; Tamè et al., 2017), we confirmed that healthy weight individuals significantly underestimate the finger length and overestimate the hand width. Notably, the distortions observed in individuals affected by obesity for both the finger length and the hand width were qualitatively and quantitatively comparable with those in the healthy weight group. Also, the increment of finger length underestimation from the thumb towards the little finger was similar in healthy weight and obesity. However, this gradient was not statistically significant in both groups. In other words, the thumb was not underestimated significantly less than the little finger. Dissimilarities with the previous experimental findings in healthy weight individuals (e.g., Longo, 2014; Longo & Haggard, 2010, 2012a) might be related to procedural differences: while in this study participants pointed the landmarks with their left index on the touch-screen monitor, in the previous studies they localised landmarks through a baton (e.g., Coelho & Gonzalez, 2018; Longo, 2014; Longo & Haggard, 2010). First, participants might have found easier to locate the landmarks directly with the index, minimising the difference in underestimation across the fingers. Moreover, it is well known that functional tools can be embodied in one’s body representation (Maravita & Iriki, 2004), shaping the representation of body metrics (Sposito et al., 2012). Thus, we cannot rule out the possibility that the use of a functional tool in previous studies influenced the experimental findings, perhaps explaining the dissimilarities with our results. In fact, tool embodiment and its possible effect on bodily representation in obesity are unproved; therefore, we explicitly avoided using a tool.

Anyway, considering that the two groups reported similar body model distortions, we may suggest that obesity did not affect body model features, as far as it concerns the hand. Indeed, the magnitude of the estimation errors was not significantly correlated with the BMI in both groups. The physiological mechanisms supporting the development of the implicit (distorted) body model in the healthy weight population, thus, might be preserved in obesity, at least as regard the hand. In fact, the magnitude of body size estimation biases might differ across body parts (Linkenauger et al., 2015; Sadibolova et al., 2019; Scarpina et al., 2017) and body surfaces (Longo & Haggard, 2012a), possibly because of dissimilar functional roles, sensorimotor features, and emotional salience. Future studies might probe this issue by investigating whether and how body model distortions might be different across several body parts.

As mentioned, the overestimation of the horizontal dimension (i.e., hand width) and the underestimation of the vertical dimension (i.e., finger length) in the landmark localisation task have been related to the presence of a higher tactile acuity on the medio-lateral axis of the hand dorsum (which is overestimated) than on the proximo-distal axis (which is underestimated), that is to anisotropies in tactile acuity across the body surface (Longo et al., 2010; Longo & Haggard, 2010). Lately, an association between tactile acuity per se and biases in more explicit body size estimation tasks has been reported (Linkenauger et al., 2015; Proske & Gandevia, 2012; Sadibolova et al., 2019), including a task (Peviani et al., 2019) that is supposed to investigate a body representation qualitatively similar but less distorted than the body model (Longo & Haggard, 2012b). Thus, considering the lower tactile acuity threshold reported in obesity than in healthy weight participants (Boles & Givens, 2011; Bussolaro et al., 2012; Falling & Mani, 2016), we probed the possible interplay between tactile acuity and body model distortions. Here, we reported no difference in tactile acuity between the two groups, neither we found a significant relationship with the BMI, in contrast with previous observations of a higher tactile acuity threshold in obesity (Boles & Givens, 2011; Bussolaro et al., 2012; Falling & Mani, 2016). However, tactile acuity seems to be specifically influenced by the body fat ratio (Boles & Givens, 2011). Thus, the lower somatosensory sensitivity previously described in obesity might be related to the fact that the body parts considered in previous studies typically have significantly higher adiposity than in healthy weight (i.e., the low back, Falling & Mani, 2016; the abdomen, Bussolaro et al., 2012). Conversely, the hand dorsum has low adiposity in both groups, possibly explaining our observation of a similar tactile acuity between obesity and healthy weight. Moreover, we might note that we specifically assessed tactile acuity in our sample, whereas previous studies (Linkenauger et al., 2015; Peviani et al., 2019; Sadibolova et al., 2019) referred to evaluations of tactile acuity across the body provided in literature. Still, alternative measures of tactile acuity may be considered (e.g., gap detection or grating orientation discrimination) because the validity of the two-point discrimination task has been questioned (see Craig & Johnson, 2000, for details).

Nevertheless, and perhaps more interestingly, we found that the hand width overestimation increases as tactile acuity decreases when considering the whole sample, supporting the hypothesis of a possible relationship between tactile acuity per se and the body model distortions in both obesity and healthy weight. This result seems in line with the previous evidence suggesting that the dimensions of body parts with high tactile acuity are estimated more accurately than those with low tactile acuity (Linkenauger et al., 2015; Sadibolova et al., 2019). Nevertheless, tactile acuity did not influence the finger length underestimation. Similarly, a differentiated effect of tactile acuity on body size estimation when considering the proximo-distal rather than medio-later axis was observed in a line length matching task (Peviani et al., 2019). As regards our study, we might speculate that as tactile acuity is higher across the hand width than length (Weber, 1996) the effect of tactile acuity on body size estimation might be stronger in this orientation. Also, we measured tactile acuity along the horizontal dimension, but if anisotropies do have a role, one might expect that length estimation might be more related to tactile acuity along the vertical dimension. Thus, according to what has been initially proposed by Longo and Haggard (2010), the specific relationship between body model distortions and anisotropies in tactile acuity across the horizontal and longitudinal body axes should be probed by future researches. Indeed, the relationship between somatosensation and body model distortions might be more complex and intricate than expected.

To sum up, our results on body model distortions in obesity seem to agree with the previous evidence about the implicit representation of body size underlying tactile distances estimation in children and adolescents (Mölbert et al., 2016) and adults (Scarpina et al., 2014). Indeed, we reported a qualitatively similar distorted implicit representation of the body in obesity and healthy weight participants, even though here we did not observe a higher magnitude of distortions in obesity. Two explanations may be given for this inconsistency. First, previous studies (Mölbert et al., 2016; Scarpina et al., 2014) investigated the implicit representation of the body metric involved in the estimation of tactile distances; conversely, in this study, we investigated the body metric underpinning position sense through the landmark localisation task. Despite these two procedures accounted for qualitatively similar distortions (i.e., an underestimation of the proximo-distal body axis and an overestimation of the medio-lateral one), the magnitude of the observed biases was reported to be different (Longo & Morcom, 2016). Second, we investigated a different part of the body, that is, the hand, whereas the previous studies focused on the forearm (Mölbert et al., 2016; Scarpina et al., 2014), the abdomen (Scarpina et al., 2014), the buttocks, the thighs, and the spine (Mölbert et al., 2016). As previously mentioned, different body surfaces have different somatosensory features (Cody et al., 2008) and emotional salience in terms of negative feelings and body dissatisfaction (Sarwer et al., 1998); thus, dissimilarities in body model distortions might be expected. Specifically, the hand might not have a high emotional salience, whereas previous studies considered body parts that are among the most unsatisfying in both obesity and healthy weight, such as the abdomen, the thigh, and the buttocks. Finally, it must be noted that Mölbert and colleagues’ (2016) work focused on children and adolescents, whose body representation is still under development and, thus, particularly malleable and unstable (Bremner et al., 2013; Kállai et al., 2017).

In the end, it is fair to note that the internal validity of the landmark localisation task to measure the body model has been questioned. First, it has been observed that qualitatively similar distortions, but different in magnitude, emerged by the localisation of both bodily and non-corporal landmarks (Saulton et al., 2015). However, these differences in magnitude disappear when correcting for a distorted visual knowledge about the hand landmarks spatial location (Saulton et al., 2017). In addition, Medina and Duckett (2017) observed that knuckles are localised as farther apart than their actual location because of a spatial bias rather than representational distortions. Thus, it has been proposed that the distortions observed in the case of the landmark localisation task might be the result of conceptual biases about the spatial location of body landmarks (Saulton et al., 2017) and/or domain-general bias in spatial localisation (Medina & Duckett, 2017), instead of being related to body representation. The truth may lie somewhere in between. Indeed, Peviani and Bottini (2020) have recently analysed the pattern of proprioceptive errors in the landmark localisation task, showing that they are suggestive of body-specific metric biases, purely proprioceptive biases, and domain-general spatial memory biases. Accordingly, even though the existence of a body model has not been questioned in these studies, future researches should further probe the adequacy of the landmark localisation task to measure the body model, as well as the possibility to use alternative tasks. For instance, when individuals match the position of the hidden knuckles or fingertips to a visual target (such as in a dynamic version of the localisation task called proprioceptive matching task), they rely on an implicit representation of the body metric similar, but slightly less distorted, than the body model (Peviani & Bottini, 2018; Peviani et al., 2020). However, the proprioceptive matching task seems not affected by the domain-general bias in spatial localisation (Peviani & Bottini, 2018) reported by Medina and Duckett (2017) in the traditional version of the task; therefore, the perceptual bias observed in the proprioceptive matching task might be more specific of the body model. Furthermore, it might be interesting to note that the results reported by Peviani and Bottini (2018) seem to support the hypothesis that the body model might be involved in both perception and action and that it can be considered a “subcomponent of the body schema,” questioning the need of a conceptual distinction between these two constructs.

To conclude, this was the first study investigating the implicit representation of bodily dimensions underpinning position sense in obesity. Overall, we reported that this implicit representation in obesity has a level of inaccuracy comparable with healthy weight individuals. Therefore, despite body representations, disorders have been often reported in obesity (Schwartz & Brownell, 2004; Weinberger et al., 2016), not all components of body representation might be affected. The characterisation of body representations in obesity indeed might be more complex than expected. However, a better characterisation of this aspect might be extremely valuable in the clinical management of this condition, and it might be crucial for the development of efficient weight-loss rehabilitative treatments.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: For this study S.T. was supported by a PhD fellowship at the Center for Mind/Brain Sciences (CIMeC), University of Trento, Italy.

ORCID iD

Sofia Tagini

References

Bell

Kirkpatrick

S. W.

Rinn

R. C.

(1986). Body image of anorexic, obese, and normal females. Journal of Clinical Psychology, 42(3), 431–439. https://doi.org/10.1002/1097-4679(198605)42:3<431::AID-JCLP2270420305>3.0.CO;2-I

Boles

D. B.

Givens

S. M.

(2011). Laterality and sex differences in tactile detection and two-point thresholds modified by body surface area and body fat ratio. Somatosensory & Motor Research, 28(3–4), 102–109.

Bremner

A. J.

Hill

E. L.

Pratt

Rigato

Spence

(2013). Bodily illusions in young children: Developmental change in visual and proprioceptive contributions to perceived hand position. PLOS ONE, 8(1), Article e51887. https://doi.org/10.1371/journal.pone.0051887

Burgess

P. R.

Wei

J. Y.

Clark

F. J.

Simon

(1982). Signaling of kinesthetic information by peripheral sensory receptors. Annual Review of Neuroscience, 5(1), 171–188. https://doi.org/10.1146/annurev.ne.05.030182.001131

Bussolaro

R. A.

Garcia

E. B.

Zanella

M. T.

Ferreira

L. M.

(2012). Impaired abdominal skin sensory function in morbid obesity and after bariatric surgery. Obesity Surgery, 22(3), 353–359. https://doi.org/10.1007/s11695-011-0485-z

Cody

F. W. J.

Garside

R. A. D.

Lloyd

Poliakoff

(2008). Tactile spatial acuity varies with site and axis in the human upper limb. Neuroscience Letters, 433(2), 103–108. https://doi.org/10.1016/j.neulet.2007.12.054

Coelho

L. A.

Gonzalez

C. L.

(2018). The visual and haptic contributions to hand perception. Psychological Research, 82(5), 866–875. https://doi.org/10.1007/s00426-017-0870-x

Cornsweet

T. N.

(1962). The staircase-method in psychophysics. The American Journal of Psychology, 75(3), 485–491. https://doi.org/10.2307/1419876

Craig

J. C.

Johnson

K. O.

(2000). The two-point threshold. Current Directions in Psychological Science, 9(1), 29–32. https://doi.org/10.1111/1467-8721.00054

10.

Dienes

(2014). Using Bayes to get the most out of non-significant results. Frontiers in Psychology, 5, Article 781.

11.

Ehrenstein

W. H.

Ehrenstein

(1999). Psychophysical methods. In Windhorst

Johansson

(Eds.), Modern techniques in neuroscience research (pp. 1211–1241). Springer. https://doi.org/10.1007/978-3-642-58552-4_43

12.

Falling

Mani

(2016). Ageing and obesity indices influences the tactile acuity of the low back regions: A cross-sectional study. Manual Therapy, 23, 25–31. https://doi.org/10.1016/j.math.2016.02.004

13.

Gallagher

(2005). How the body shapes the mind. Clarendon Press.

14.

Ganea

Longo

M. R.

(2017). Projecting the self outside the body: Body representations underlying proprioceptive imagery. Cognition, 162, 41–47. https://doi.org/10.1016/j.cognition.2017.01.021

15.

Green

B. G.

(1982). The perception of distance and location. Perception & Psychophysics, 31(4), 315–323.

16.

Head

Holmes

(1911). Sensory disturbances from cerebral lesions. Brain, 34(2–3), 102–254. https://doi.org/10.1093/brain/34.2-3.102

17.

JASP Team (2019). JASP (Version 0.11.1).

18.

Jeffreys

(1961). Theory of probability. Oxford University Press.

19.

Johnstone

A. M.

Stewart

A. D.

Benson

P. J.

Kalafati

Rectenwald

Horgan

(2008). Assessment of body image in obesity using a digital morphing technique. Journal of Human Nutrition and Dietetics, 21(3), 256–267. https://doi.org/10.1111/j.1365-277X.2008.00862.x

20.

Kállai

Kincses

Lábadi

Dorn

Szolcsányi

Darnai

Hupuczi

Janszky

Csathó

Á.

Kincses

Lábadi

Dorn

Szolcsányi

Darnai

Hupuczi

Janszky

Rpád Csathó

. (2017). Multisensory integration and age-dependent sensitivity to body representation modification induced by the rubber hand illusion. Cognitive Processing, 18(4), 349–357. https://doi.org/10.1007/s10339-017-0827-4

21.

Keizer

Smeets

M. A. M.

Dijkerman

H. C.

Uzunbajakau

S. A.

van Elburg

Postma

(2013). Too fat to fit through the door: First evidence for disturbed body-scaled action in anorexia nervosa during locomotion. PLOS ONE, 8(5), Article e64602. https://doi.org/10.1371/journal.pone.0064602

22.

Klein

S. A.

(2001). Measuring, estimating, and understanding the psychometric function: A commentary. Perception & Psychophysics, 63(8), 1421–1455. https://link-springer-com-443.web.bisu.edu.cn/content/pdf/10.3758/BF03194552.pdf

23.

Linkenauger

S. A.

Wong

H. Y.

Geuss

Stefanucci

J. K.

McCulloch

K. C.

Bülthoff

H. H.

Mohler

B. J.

Proffitt

D. R.

(2015). The perceptual homunculus: The perception of the relative proportions of the human body. Journal of Experimental Psychology: General, 144(1), 103–113. https://doi.org/10.1037/xge0000028

24.

Longo

M. R.

(2014). The effects of immediate vision on implicit hand maps. Experimental Brain Research, 232(4), 1241–1247. https://doi.org/10.1007/s00221-014-3840-1

25.

Longo

M. R.

(2015a). Implicit and explicit body representations. European Psychologist, 20(1), 6–15. https://doi.org/10.1027/1016-9040/a000198

26.

Longo

M. R.

(2015b). Posture modulates implicit hand maps. Consciousness and Cognition, 36, 96–102. https://doi.org/10.1016/j.concog.2015.06.009

27.

Longo

M. R.

(2015c). Types of body representation. In Fischer

Coello

(Eds.), Perceptual and emotional embodiment (pp. 125–142). Routledge.

28.

Longo

M. R.

(2017). Expansion of perceptual body maps near—But not across—The wrist. Frontiers in Human Neuroscience, 11, Article 111. https://doi.org/10.3389/fnhum.2017.00111

29.

Longo

M. R.

(2018). The effects of instrumental action on perceptual hand maps. Experimental Brain Research, 236(11), 3113–3119. https://doi.org/10.1007/s00221-018-5360-x

30.

Longo

M. R.

Azañón

Haggard

(2010). More than skin deep: Body representation beyond primary somatosensory cortex. Neuropsychologia, 48(3), 655–668. https://doi.org/10.1016/j.neuropsychologia.2009.08.022

31.

Longo

M. R.

Ghosh

Yahya

(2015). Bilateral symmetry of distortions of tactile size perception. Perception, 44(11), 1251–1262. https://doi.org/10.1177/0301006615594949

32.

Longo

M. R.

Haggard

(2010). An implicit body representation underlying human position sense. Proceedings of the National Academy of Sciences, 107(26), 11727–11732. https://doi.org/10.1073/pnas.1003483107

33.

Longo

M. R.

Haggard

(2011). Weber’s illusion and body shape: Anisotropy of tactile size perception on the hand. Journal of Experimental Psychology: Human Perception and Performance, 37(3), 720–726. https://doi.org/10.1037/a0021921

34.

Longo

M. R.

Haggard

(2012a). A 2.5-D representation of the human hand. Journal of Experimental Psychology: Human Perception and Performance, 38(1), 9–13. https://doi.org/10.1037/a0025428

35.

Longo

M. R.

Haggard

(2012b). Implicit body representations and the conscious body image. Acta Psychologica, 141(2), 164–168. https://doi.org/10.1016/j.actpsy.2012.07.015

36.

Longo

M. R.

Long

Haggard

(2012). Mapping the invisible hand. Psychological Science, 23(7), 740–742. https://doi.org/10.1177/0956797612441219

37.

Longo

M. R.

Lulciuc

Sotakova

(2019). No evidence of tactile distance anisotropy on the belly. Royal Society Open Science, 6(3), Article 180866. https://doi.org/10.1098/rsos.180866

38.

Longo

M. R.

Morcom

(2016). No correlation between distorted body representations underlying tactile distance perception and position sense. Frontiers in Human Neuroscience, 10, Article 593. https://doi.org/10.3389/fnhum.2016.00593

39.

Longo

M. R.

Sadibolova

(2013). Seeing the body distorts tactile size perception. Cognition, 126(3), 475–481. https://doi.org/10.1016/j.cognition.2012.11.013

40.

Verhagen

Wagenmakers

E.-J.

(2016). Harold Jeffreys’s default Bayes factor hypothesis tests: Explanation, extension, and application in psychology. Journal of Mathematical Psychology, 72, 19–32. https://doi.org/10.1016/j.jmp.2015.06.004

41.

Mancini

Bauleo

Cole

Lui

Porro

C. A.

Haggard

Iannetti

G. D.

(2014). Whole-body mapping of spatial acuity for pain and touch. Annals of Neurology, 75(6), 917–924. https://doi.org/10.1002/ana.24179

42.

Maravita

Iriki

(2004). Tools for the body (schema). Trends in Cognitive Sciences, 8(2), 79–86. https://doi.org/10.1016/j.tics.2003.12.008

43.

Marks

L. E.

Girvin

J. P.

Quest

D. O.

Antunes

J. L.

Ning

O’Keefe

M. D.

Dobelle

W. H.

(1982). Electrocutaneous stimulation II. The estimation of distance between two points. Perception & Psychophysics, 32(6), 529–536. https://doi.org/10.3758/BF03204206

44.

Mattioni

Longo

M. R.

(2014). The effects of verbal cueing on implicit hand maps. Acta Psychologica, 153, 60–65. https://doi.org/10.1016/j.actpsy.2014.09.009

45.

Medina

Duckett

(2017). Domain-general biases in spatial localization: Evidence against a distorted body model hypothesis. Journal of Experimental Psychology: Human Perception and Performance, 43(7), 1430–1443. https://doi.org/10.1037/xhp0000397

46.

Mölbert

S. C.

Sauer

Dammann

Zipfel

Teufel

Junne

Enck

Giel

K. E.

Mack

(2016). Multimodal body representation of obese children and adolescents before and after weight-loss treatment in comparison to normal-Weight children. PLOS ONE, 11(11), Article e0166826. https://doi.org/10.1371/journal.pone.0166826

47.

Paillard

(1999). Body schema and body image—A double dissociation. Motor Control, Today and Tomorrow, 48(3), 197–214. https://doi.org/10.1016/j.neuropsychologia.2009.09.022

48.

Peviani

Bottini

(2018). The distorted hand metric representation serves both perception and action. Journal of Cognitive Psychology, 30(8), 880–893. https://doi.org/10.1080/20445911.2018.1538154

49.

Peviani

Bottini

(2020). Proprioceptive errors in the localization of hand landmarks: What can be learnt about the hand metric representation? PLOS ONE, 15(7), Article e0236416. https://doi.org/10.1371/journal.pone.0236416

50.

Peviani

Liotta

Bottini

(2020). The motor system (partially) deceives body representation biases in absence of visual correcting cues. Acta Psychologica, 203, Article 103003. https://doi.org/10.1016/j.actpsy.2020.103003

51.

Peviani

Melloni

Bottini

(2019). Visual and somatosensory information contribute to distortions of the body model. Scientific Reports, 9(1), Article 13570. https://doi.org/10.1038/s41598-019-49979-0

52.

Piryankova

I. V.

Wong

H. Y.

Linkenauger

S. A.

Stinson

Longo

M. R.

Bülthoff

H. H.

Mohler

B. J.

(2014). Owning an overweight or underweight body: Distinguishing the physical, experienced and virtual body. PLOS ONE, 9(8), Article e103428. https://doi.org/10.1371/journal.pone.0103428

53.

Proske

Gandevia

S. C.

(2012). The proprioceptive senses: Their roles in signaling body shape, body position and movement, and muscle force. Physiological Reviews, 92(4), 1651–1697. https://doi.org/10.1152/physrev.00048.2011

54.

Rouder

J. N.

Speckman

P. L.

Sun

Morey

R. D.

Iverson

(2009). Bayesian t tests for accepting and rejecting the null hypothesis. Psychonomic Bulletin & Review, 16(2), 225–237. https://doi.org/10.3758/PBR.16.2.225

55.

Sadibolova

Ferrè

E. R.

Linkenauger

S. A.

Longo

M. R.

(2019). Distortions of perceived volume and length of body parts. Cortex, 111, 74–86. https://doi.org/10.1016/j.cortex.2018.10.016

56.

Sarwer

D. B.

Wadden

T. A.

Foster

G. D.

(1998). Assessment of body image dissatisfaction in obese women: Specificity, severity, and clinical significance. Journal of Consulting and Clinical Psychology, 66(4), 651–654. https://doi.org/10.1037/0022-006X.66.4.651

57.

Saulton

Bülthoff

H. H.

De la Rosa

(2017). Conceptual biases explain distortion differences between hand and objects in localization tasks. Journal of Experimental Psychology: Human Perception and Performance, 43(7), 1444–1453. https://doi.org/10.1037/xhp0000396

58.

Saulton

Dodds

T. J.

Bülthoff

H. H.

De la Rosa

(2015). Objects exhibit body model like shape distortions. Experimental Brain Research, 233, 1471–1479. https://doi.org/10.1007/s00221-015-4221-0

59.

Saulton

Longo

M. R.

Wong

H. Y.

Bülthoff

H. H.

de la Rosa

(2016). The role of visual similarity and memory in body model distortions. Acta Psychologica, 164, 103–111. https://doi.org/10.1016/j.actpsy.2015.12.013

60.

Scarpina

Castelnuovo

Molinari

(2014). Tactile mental body parts representation in obesity. Psychiatry Research, 220(3), 960–969. https://doi.org/10.1016/j.psychres.2014.08.020

61.

Scarpina

Cau

Cimolin

Galli

Castelnuovo

Priano

Pianta

Corti

Mauro

Capodaglio

(2017). Body-scaled action in obesity during locomotion: Insights on the nature and extent of body representation disturbances. Journal of Psychosomatic Research, 102, 34–40. https://doi.org/10.1016/j.jpsychores.2017.09.004

62.

Scarpina

Serino

Keizer

Chirico

Scacchi

Castelnuovo

Mauro

Riva

(2019). The effect of a virtual-reality full-body illusion on body representation in obesity. Journal of Clinical Medicine, 8(9), Article 1330. https://doi.org/10.3390/jcm8091330

63.

Schwartz

M. B.

Brownell

K. D.

(2004). Obesity and body image. Body Image, 1(1), 43–56. https://doi.org/10.1016/S1740-1445(03)00007-X

64.

Spitoni

G. F.

Galati

Antonucci

Haggard

Pizzamiglio

(2010). Two forms of touch perception in the human brain. Experimental Brain Research, 207(3–4), 185–195. https://doi.org/10.1007/s00221-010-2446-5

65.

Sposito

Bolognini

Vallar

Maravita

(2012). Extension of perceived arm length following tool-use: Clues to plasticity of body metrics. Neuropsychologia, 50(9), 2187–2194. https://doi.org/10.1016/j.neuropsychologia.2012.05.022

66.

Tamè

Bumpus

Linkenauger

S. A.

Longo

M. R.

(2017). Distorted body representations are robust to differences in experimental instructions. Attention, Perception, and Psychophysics, 79(4), 1204–1216. https://doi.org/10.3758/s13414-017-1301-1

67.

Thomas

(1997). Retrospective power analysis. Conservation Biology, 11(1), 276–280. https://doi.org/10.1046/j.1523-1739.1997.96102.x

68.

van Beers

R. J.

Sittig

A. C.

Denier van der Gon

J. J

. (1998). The precision of proprioceptive position sense. Experimental Brain Research, 122(4), 367–377. https://doi.org/10.1007/s002210050525

69.

Weber

E. H.

(1996). EH Weber on the tactile senses. Psychology Press.

70.

Weinberger

N. A.

Kersting

Riedel-Heller

S. G.

Luck-Sikorski

(2016). Body dissatisfaction in individuals with obesity compared to normal-weight individuals: A systematic review and meta-analysis. Public Health, 9(6), 424–441. https://doi.org/10.1159/000454837

71.

World Health Organization. (2020). Obesity. https://www.who.int/topics/obesity/en/