A grating orientation task with controlled dome elevation speed and contact force: stimulus conditions and day-to-day reliability Authors: Kei Saito, Sho Kojima, Taiga Hiraga, Yusuke Sekine, Hideaki Onishi Published: 2026 ABSTRACT AbstractThe grating orientation discrimination task (GOT) is typically used to assess tactile spatial acuity. However, performance is influenced by stimulus speed, contact force between the stimulator and the skin, and the method of discrimination. This study aimed to explore suitable stimulus conditions for a novel GOT approach and evaluate its feasibility, day-to-day reliability, and psychometric properties for assessing tactile spatial acuity.•Tested different stimulus speeds and contact forces.•Compared a novel three-alternative method with the conventional approach.•Assessed reliability and psychometric properties of GOT measurements.Keywords: Grating orientation discrimination, Tactile spatial acuity, Lifting speed, Contact force, Reproducibility Method overview In addition, among 14 neurologically healthy participants, GOT was performed twice ≥1 day apart on the right index finger under varying stimulus conditions. The novel method, using a hemispherical dome elevation speed of 15 mm/s and a contact force of 4 N, yielded GOT thresholds comparable to those obtained from historical data using the conventional two-alternative method in 40 participants. Furthermore, preliminary evidence indicated acceptable reproducibility and favorable psychometric curve fitting. These findings suggest that the novel GOT approach may provide a feasible method for assessing tactile spatial acuity, yielding thresholds comparable to those obtained with the conventional method and providing preliminary evidence of acceptable reliability and improved psychometric curve fitting. Specifications table Subject areaNeuroscienceMore specific subject areaTactile perceptionName of your methodGrating Orientation discriminationName and reference of original methodResource availability Background The grating orientation discrimination task (GOT) is widely used to assess tactile spatial acuity, which varies across body parts (fingers: 0.98 mm, tongue: 0.64 mm, and lips: 0.55 mm) [1]. It requires participants to discriminate the orientation of grooved stimuli applied to body parts, including fingers and hands [2]. However, tactile spatial acuity is influenced by individual characteristics. For instance, the GOT threshold is reportedly lower in women than in men and individuals with small hands have a relatively low threshold, regardless of sex [3]. Further, the GOT threshold increases with age [4] and can be modulated by peripheral electrical stimulation [5], transcranial electrical stimulation [[6], [7], [8], [9], [10], [11], [12], [13], [14], [15]], and tactile perception training [[16], [17], [18]]. Recent tactile psychophysics studies continue to use grating orientation and spatial discrimination paradigms to assess tactile spatial acuity and tactile sensory function [[19], [20], [21], [22]]. Thus, the GOT threshold is considered a useful indicator of changes in tactile perception. In conventional GOT measurements using a constant stimulus [[6], [7], [8], [9], [10], [11], [12], [13], [14], [15]], participants judged groove orientation using a two-alternative forced-choice method after being shown a series of eight domes with groove widths ranging from the largest (3.0 mm) to the smallest (0.35 mm) [5,[12], [13], [14], [15],18]. However, the correct response rate for groove widths below the discrimination threshold approaches chance levels, potentially increasing variability at each groove width, resulting in poorer fitting of the psychometric function [5]. In a two-alternative forced-choice paradigm, participants must select one of two orientations even when uncertain, which may introduce guessing-related variability, particularly near threshold levels. In contrast, allowing participants to report uncertainty may reduce the influence of guessing and improve distinction between perceived orientation discrimination and uncertain responses. Consistent with this idea, Yokota et al. (2020) showed stable and reliable two-point discrimination thresholds when participants were asked to report their perception of two points only if certain [23]. Participants were instructed to report two points only when confident and one point otherwise (including when unsure). These findings suggest that incorporating an uncertainty response option into the GOT paradigm may produce a more consistent stimulus–response relationship and improve psychometric curve fitting for threshold estimation. Tactile spatial acuity is influenced by stimulus conditions. In previous research, the tactile discrimination threshold was modulated by stimulus pin speed [23] and the contact force between the stimulus surface and skin [4]. Accordingly, GOT threshold reproducibility may be affected by the stimulus speed of the hemispherical dome with eight different groove widths and the contact force between the hemispherical dome and skin. Thus, this study aimed to characterize the influence of stimulus speed and contact force on the GOT threshold based on its intraclass correlation coefficient (ICC) and Bland–Altman plots and determine the optimal stimulus speed and contact force for day-to-day reliability. In addition, we evaluated the day-to-day reliability, threshold comparability, and psychometric curve-fitting performance of a novel GOT measurement method using the selected hemispherical dome elevation speed and contact force. Methods Participants This study included 14 right-handed adults who underwent GOT measurement using a three-alternative forced-choice method (novel GOT measurement) and 40 right-handed adults who underwent GOT measurement using a two-alternative forced-choice method (conventional GOT measurement). The novel GOT measurement group (8 men, 6 women) had a mean ± SD age of 21.4 ± 0.6 years (range, 21–23 years). The conventional GOT measurement group (29 men, 11 women) had a mean ± SD age of 22.2 ± 2.0 years (range, 20–32 years). Data from the conventional GOT measurements were previously collected as part of a published study on the effects of peripheral and transcranial electrical stimulation on tactile discrimination acuity [5,14]. This study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Niigata University of Health and Welfare (19392–240912, September 25, 2024). All participants provided written informed consent. Participants were informed about the general procedures of the study but were blinded to the specific study hypotheses. Experimental design A schematic diagram of the novel GOT measurement method is shown in Fig. 1. This study comprised two experiments. The first involved three stimulus speed conditions (i.e., 10, 15, and 20 mm/s) using the hemispherical dome with eight different groove widths. In the second experiment, the contact force between the hemispherical dome and the skin was varied (i.e., 4 N, 7 N, and 10 N). Both stimulus speed and contact force sessions were performed in a random order on the same day with a 5-min interval (test 1) and ≥1 day later (test 2). To minimize order effects, GOT tasks were randomized separately for each participant. A random allocation list was created in Microsoft Excel before testing to determine task allocation. Schematic representation of the GOT method to investigate the effect of stimulus speed and pressure on tactile spatial acuity judgments. Tactile spatial acuity was assessed by measuring the orientation discrimination of gratings of varying sizes (0.35–3.0 mm) through a three-alternative forced-choice task. In experiment 1, GOT measurements were performed randomly using three tactile stimuli elevation speeds from eight hemispherical domes. In experiment 2, GOT measurements were performed in a random order under three contact-force conditions using eight hemispherical domes. Experiments 1 and 2 were conducted ≥ 1 day apart to provide a washout period. GOT, grating orientation discrimination task Grating orientation discrimination task To assess the tactile discrimination acuity of the right index finger, in the novel GOT measurement, 16 stimuli were randomly applied to the right index finger using a hemispherical dome with eight different groove widths (3.0, 2.0, 1.5, 1.2, 1.0, 0.75, 0.5, and 0.35 mm; total of 8 stimulus widths × 16 times = 128 dome presentations) and a custom-made device, which automatically raised and lowered the hemispherical dome (Tec Gihan Co., Ltd., Kyoto, Japan). The novel GOT device allowed control of the contact force applied to the finger; the pressure sensor was calibrated before each experimental session to ensure accurate application of the target contact force during testing. Participants were asked to judge the dome orientation (i.e., orthogonal or parallel direction) relative to the long axis of the index finger using a three-alternative forced-choice method; “no judgment,” was the required response if they were unsure of the dome orientation. In experiment 1, the hemispherical dome was pressed against the right index finger at 7 N for 1 s and the elevation speed was set to 10, 15, or 20 mm/s. In experiment 2, the hemispherical dome was pressed on the right index finger for 1 s at 15 mm/s elevation speed, which showed the highest ICC in experiment 1. The contact force between the hemispherical dome and the right index finger was set to 4, 7, or 10 N. In both experiments 1 and 2, to minimize fatigue, participants had 5-min rest periods between tasks; fatigue was assessed verbally as needed during the experiment. No feedback regarding response accuracy was provided during GOT tasks to minimize explicit learning effects across repeated trials. For conventional GOT measurements, 16 stimuli from a hemispherical dome with eight different groove widths (3.0, 2.0, 1.5, 1.2, 1.0, 0.75, 0.5, and 0.35 mm) were serially applied to the right index finger in decreasing width order (from 3.0 mm to 0.35 mm) using a custom-made device (S-16,026; Takei Scientific Instruments Co. Ltd., Niigata, Japan). Participants were asked to judge the dome orientation (orthogonal or parallel direction) relative to the long axis of their index finger using a two-alternative forced choice method, as previously published [5,14]. GOT was conducted in a comfortable laboratory environment. Room temperature was not fixed, humidity was not specifically controlled, and no standardized skin preparation procedure was performed before GOT measurements. Data analysis Data were analyzed using MATLAB 2020a (The Mathworks Inc., Natick, MA, United States). The percentage of correct answers was determined for each stimulus width; any "no judgment" responses for a given stimulus width were considered incorrect because the GOT threshold was intended to reflect the ability to correctly identify the orientation of the tactile stimulus. Thus, a response indicating that the participant could not judge the stimulus orientation was considered a failure to perform the discrimination task. Psychometric curves were fitted using a logistic regression model using the glmfit function in MATLAB [5,[12], [13], [14], [15]]. For each participant and session, the proportion of correct responses at each groove width was fitted using a binomial distribution with a logit link function. The number of trials per groove width was used as binomial denominator. A fitted psychometric function was then generated using the MATLAB glmval function. GOT thresholds were the groove widths corresponding to predetermined target response levels (50% for the novel method and 75% for the conventional method) on the fitted logistic function. Specifically, thresholds were calculated using the estimated intercept and slope of the logistic model as follows: where p indicates the target response level, and β0 and β1 indicate the intercept and slope of the fitted logistic model, respectively. Repeated measurements across sessions were fitted separately and the resulting session-specific threshold estimates used for reliability analysis. In addition, to assess the logistic regression fit for each correct response rate, the adjusted coefficient of determination was calculated using the following equation: where Y¯ was the mean value of Yi, n was the number of groove widths, and d was the degree of the regression equation. Statistical analysis All statistical analyses were performed using SPSS (version 25, IBM Corp., Armonk, NY, USA). To assess the test-retest reliability of the GOT threshold between tests 1 and 2, ICCs (one-way random effects, absolute agreement, and single-measure) were calculated and interpreted as follows: poor (<0.5), moderate (0.5–0.75), good (0.76–0.9), or excellent (>0.9) [24]. Bland–Altman plots were generated to assess test–retest agreement. The repeatability coefficient (RC) was calculated as 1.96 times the standard deviation of the test–retest differences, and the 95% limits of agreement were defined as the mean difference ± RC [25]. Intertest differences were evaluated using a one-sample t-test. For each test, linear regression analysis was used to evaluate the relationship between the differences between tests 1 and 2 and the mean of repeated tests. To assess test-retest reliability and psychometric curve-fitting performance of a novel GOT measurement method, the GOT threshold, ICC, and R2adjusted value were compared with those of the conventional GOT measurement method. The difference in GOT threshold and the R²-adjusted value between novel and conventional GOT measurement methods was analyzed using an independent-sample t-test. Confidence intervals for individual GOT threshold estimates were not calculated. Statistical significance was set at p < 0.05. Method validation Test-retest reliability In experiment 1, the day-to-day reliability of GOT thresholds was moderate to good across all hemispherical dome elevation speeds, with ICC(1,1) ranging from 0.678 to 0.884 (Fig. 2A). Specifically, the ICC(1,1) for an elevation speed of 10 mm/s was 0.678, with a 95% confidence interval (CI) of 0.267–0.883; that for an elevation speed of 15 mm/s was 0.884, with a 95% CI of 0.687–0.961; and that for an elevation speed of 20 mm/s was 0.704, with a 95% CI of 0.312–0.893. Among all tested elevation speeds, the ICC(1,1) point estimate for the GOT threshold was highest at 15 mm/s (ICC(1,1) = 0.884), compared with 10 mm/s (ICC(1,1) = 0.678) and 20 mm/s (ICC(1,1) = 0.704). However, given the relatively wide confidence intervals and modest sample size, differences in ICC estimates between elevation speeds should be interpreted cautiously. Day-to-day agreement across all elevation speeds was good (Fig. 2B). No significant differences between tests 1 and 2 were observed for all GOT measurements (10 mm/s, p = 0.343; 15 mm/s, p = 0.291; and 20 mm/s, p = 0.911). The association between the differences observed in tests 1 and 2 and the mean GOT thresholds was not significant across GOT measurements (10 mm/s, p = 0.411; 15 mm/s, p = 0.669; 20 mm/s, p = 0.082). The elevation speed of tactile stimuli from eight hemispherical domes modulates GOT thresholds. (A) Day-to-day reliability of GOT thresholds as assessed using the ICC. ICC(1,1) values with 95% CIs: 10 mm/s, ICC(1,1) = 0.678, 95% CI = 0.267–0.883; 15 mm/s, ICC(1,1) = 0.884, 95% CI = 0.687–0.961; and 20 mm/s, ICC(1,1) = 0.704, 95% CI = 0.312–0.893. (B) Bland–Altman plots comparing the results between tests 1 and 2 for elevation speeds of 10, 15, and 20 mm/s. Solid line, the mean bias; dashed lines, lower and upper limits of agreement, calculated as mean bias ± 1.96 SD of differences. The mean GOT threshold is plotted on the x-axis and the GOT threshold difference between tests 1 and 2 on the y-axis. For the 10 mm/s elevation speed, the Bland–Altman analysis showed a mean bias of 0.050 mm, with lower and upper limits of agreement of −0.324 and 0.424 mm, respectively. For the 15 mm/s elevation speed, the Bland–Altman analysis showed a mean bias of 0.035 mm, with lower and upper limits of agreement of −0.201 and 0.272 mm, respectively. For the 20 mm/s elevation speed, the Bland–Altman analysis showed a mean bias of −0.006 mm, with lower and upper limits of agreement of −0.401 and 0.388 mm, respectively. ICC, intraclass correlation coefficient: CI, confidence interval; GOT, grating orientation discrimination task In experiment 2, the day-to-day reliability of the GOT thresholds was good or excellent for all contact forces, with ICC(1,1) ranging from 0.808 to 0.916 (Fig. 3A). Specifically, the ICC(1,1) for 4 N contact force was 0.916, with a 95% CI of 0.767–0.972; 0.907 for the 7 N contact force, with a 95% CI of 0.744–0.969; and 0.808 for the 10 N contact force, with a 95% CI of 0.515–0.933. Among all tested contact forces, the ICC(1,1) point estimate for the GOT threshold was highest at 4 N (ICC(1,1) = 0.916), followed by 7 N (ICC(1,1) = 0.907) and 10 N (ICC(1,1) = 0.808). Although both 4 and 7 N conditions showed high ICC point estimates, differences between contact force conditions should be interpreted cautiously because of the relatively wide confidence intervals and modest sample size. Day-to-day agreement among all contact forces was good (Fig. 3B). No significant differences were observed between tests 1 and 2 across GOT measurements (4 N, p = 0.092; 7 N, p = 0.831; 10 N, p = 0.802; one-sample t-test). Additionally, there was no significant association between test-retest differences and mean GOT thresholds for any contact force (4 N, p = 0.194; 7 N, p = 0.195; 10 N, p = 0.497). Tactile stimuli force from eight hemispherical domes modulated the GOT thresholds. (A) Day-to-day reliability of GOT thresholds as assessed using the ICC. ICC(1,1) values with 95% CIs: 4 N, ICC(1,1) = 0.916, 95% CI = 0.767–0.972; 7 N, ICC(1,1) = 0.907, 95% CI = 0.744–0.969; and 10 N, ICC(1,1) = 0.808, 95% CI = 0.515–0.933. (B) Bland–Altman plots comparing the results between tests 1 and 2 for 4, 7, and 10 N. Solid line, mean bias; dashed lines, lower and upper limits of agreement, calculated as mean bias ± 1.96 SD of differences. For a 4 N contact force, the Bland–Altman analysis showed a mean bias of 0.052 mm, with lower and upper limits of agreement of −0.156 and 0.259 mm, respectively. For a 7 N contact force, the Bland–Altman analysis showed a mean bias of 0.007 mm, with lower and upper limits of agreement of −0.219 and 0.232 mm, respectively. For a 10 N contact force, the Bland–Altman analysis showed a mean bias of 0.011 mm, with lower and upper limits of agreement of −0.306 and 0.328 mm, respectively. ICC, intraclass correlation coefficient: CI, confidence interval; GOT, grating orientation discrimination task Comparison between novel and conventional GOT measurement methods The GOT threshold calculated using the novel method (ICC(1,1) = 0.916, with a 95% CI of 0.767–0.972) had a higher relative reliability than that calculated using the conventional method (ICC(1,1) = 0.633, with a 95% CI of 0.406–0.787; Fig. 4A). Compared with the conventional method, the novel method had a similar GOT discrimination threshold (t(52) = 0.245, p = 0.807, d = 0.076; Fig. 4C) and a significantly higher coefficient of determination (t(52) = 8.348, p = 3.60e–11, d = 2.592). Comparison of GOT thresholds, reproducibility, and psychometric curve fitting between the novel and conventional methods. Test-retest reliability of GOT thresholds using novel and conventional GOT methods. The interclass correlation coefficient for the novel GOT measurement is the same as that in Fig. 3. (B) Bland–Altman plots comparing results between tests 1 and 2 for novel and conventional GOT methods. Bias (solid line) and limits of agreement (dotted line) are shown. The mean GOT threshold is plotted on the x-axis and the GOT threshold difference between tests 1 and 2 on the y-axis. Bland–Altman plots for the novel GOT measurement in the left panel are the same as those in Fig. 3. (C) There was no significant difference in GOT thresholds between novel and conventional GOT methods (t(52) = 0.245, p = 0.807, d = 0.076). The coefficient of determination obtained with the novel GOT method was significantly higher than with the conventional GOT method (t(52) = 8.348, p = 3.60e-11, d = 2.592). **p < 0.01. Limitations Several limitations should be acknowledged. First, the comparison between the novel and conventional GOT methods was based on historical control data from previous studies. Although both novel and conventional GOT datasets were collected in the same laboratory by the same researcher, the comparison was based on historical control data from previously published studies [5,14]. Therefore, the datasets were not collected concurrently nor obtained from the same participants. In addition, some exclusion criteria differed because of methodological differences and threshold definitions. Thus, potential confounding factors related to participant characteristics, temporal effects, calibration conditions, and minor procedural differences cannot be completely excluded. Accordingly, the comparison with the conventional method should be interpreted cautiously, and future studies should directly compare the novel and conventional GOT methods within the same participants under identical experimental conditions. Second, the interpretation of threshold estimates may be influenced by the treatment of “no judgment” responses. In this study, “no judgment” responses were considered incorrect responses as they indicated that the participant could not identify the stimulus orientation. However, this may have lowered the estimated proportion of correct responses, particularly at more difficult groove widths, thereby influencing threshold estimation. Alternative approaches, such as excluding “no judgment” trials, assigning probabilistic weights, or modeling uncertainty responses as a separate response category, may yield different threshold estimates and should be examined in future studies. Third, the generalizability of the present findings is limited. All participants were young, healthy, right-handed adults. Therefore, the applicability of the proposed GOT method to older adults, left-handed individuals, patients with neurological disorders, and clinical populations with tactile sensory impairments remains unclear. Future studies should examine whether the present findings can be replicated in more diverse populations and clinical groups for whom tactile spatial acuity assessments are particularly relevant. Despite our efforts to minimize fatigue and learning effects through rest periods, verbal fatigue checks and withholding feedback on response accuracy, potential learning habituation or fatigue effects remain a concern. This is because tactile discrimination testing was repeated across multiple stimulus conditions and sessions. In addition, room temperature was not fixed, humidity not controlled, and no standardized skin preparation procedure was performed before GOT measurements. Future studies should consider a more rigorous control of environmental and skin conditions during tactile threshold assessments. CRediT author statement Kei Saito: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Writing–Original Draft, Visualization, Supervision, Project administration, and Funding acquisition. Sho Kojima: Formal analysis, and Writing–Review. Taiga Hiraga: Investigation. Yusuke Sekine: Investigation. Hideaki Onishi: Writing–Review & Editing, Supervision, and Funding acquisition. All authors have read and approved the final version of the submitted manuscript. Ethics statements This study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of Niigata University of Health and Welfare. The ethics approval number is 19392–240912, dated September 25, 2024. All participants provided written informed consent. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments The authors would like to thank all participants for their cooperation. The authors would like to thank Enago (www.enago.jp) for their English language review. This work was supported by Grants-in-Aid for Scientific Research (C) from the Japan Society for the Promotion of Science (JSPS) [grant numbers 20K11194 and 23K10490] and a Grant-in-Aid for Scientific Research (A) from the JSPS [grant number 19H0109]. Data availability Data will be made available on request. References Associated Data Data Availability Statement Data will be made available on request.