A Comparative Analysis of Smartphone and Standard Tools for Touch Perception Assessment Across Multiple Body Sites

TL;DR

This work addresses the need for high-resolution, autonomous vibrotactile assessment beyond traditional clinical tools. It introduces an iOS smartphone platform that uses Core Haptics and a one-up/one-down staircase to measure vibration perception thresholds (VPT) across six body sites, comparing results with tuning fork and monofilament tests in healthy adults. The smartphone approach reproduced known anatomical and age-related trends and showed moderate correlation with tuning fork measurements, with weaker associations to monofilaments overall (strengthened when focusing on index finger and big toe). These findings indicate feasibility for smartphone-based, remote monitoring of vibrotactile function and potential future use in disease progression tracking, while highlighting the need to validate across more devices and diverse populations.

Abstract

Tactile perception plays an important role in activities of daily living, and it can be impaired in individuals with certain medical conditions. The most common tools used to assess tactile sensation, the Semmes-Weinstein monofilaments and the 128 Hz tuning fork, have poor repeatability and resolution. Long term, we aim to provide a repeatable, high-resolution testing platform that can be used to assess vibrotactile perception through smartphones without the need for an experimenter to be present to conduct the test. We present a smartphone-based vibration perception measurement platform and compare its performance to measurements from standard monofilament and tuning fork tests. We conducted a user study with 36 healthy adults in which we tested each tool on the hand, wrist, and foot, to assess how well our smartphone-based vibration perception thresholds (VPTs) detect known trends obtained from standard tests. The smartphone platform detected statistically significant changes in VPT between the index finger and foot and also between the feet of younger adults and older adults. Our smartphone-based VPT had a moderate correlation to tuning fork-based VPT. Our overarching objective is to develop an accessible smartphone-based platform that can eventually be used to measure disease progression and regression.

Figures (6)

• Figure 1: (A) Filtered smartphone vibration acceleration data for hapticIntensities of 0.15 and 0.25 using the setup described in detail in adenekan2022feasibility and yoshida2023cognitive, with the slight modification that the phone was placed on a pillow, not held in the hand. (B) Peak accelerations of the waveforms for each hapticIntensity (indicated with pink points in (A)). (C) Sample perception data for one trial of the smartphone VPT exam showcasing the staircase method.
• Figure 2: (A) The six body locations were tested with all three measurement methods. The smartphone was placed on a pillow and participants were asked to contact the phone as shown in this figure. During the monofilament and tuning fork tests, participants sat in a chair, and a pillow resting on a desk or coffee table allowed participants to comfortably support their limbs. (B) The filtered vibration waveforms for the six configurations shown in (A). For these measurements hapticIntensity was set to 0.25 and hapticSharpness was set to 1.
• Figure 3: Perception thresholds at the index finger and the big toe for measurement method. Individual thresholds (left) as well as group mean and standard deviation (right) are shown for the smartphone (A) and tuning fork (B). Individual thresholds (left) as well as group median and 25th and 75th quantiles (right) are shown for the monofilaments (C). All three modalities detect statistically significantly poorer perceptual resolution at the big toe than at the index finger.
• Figure 4: Perception thresholds in younger adults (left) and older adults (right) at the index finger (top) and big toe (bottom). Individual thresholds as well as group mean and standard deviation are shown for the smartphone (A) and tuning fork (B). Individual thresholds as well as median and 25th and 75th quantiles are shown for the monofilaments (C).
• Figure 5: Correlations between each of the measurement methods. Pearson's correlation coefficient is shown for the smartphone and tuning fork data (A). Spearman's correlation coefficient is shown for correlations between the smartphone and monofilaments data (B) and between the monofilaments and tuning fork data (C). As a reminder, (B) and (C) do not have best fit lines because Spearman's correlation is used to describe monotonic, not necessarily linear relationships. (D), (E), and (F) are the corresponding figures for (A), (B), and (C), when only the index finger (a convenient location for perception testing) and big toe (a clinically relevant location) are included.