This cross-sectional pilot study was designed to examine the characteristics of the tracing task and its potential use in clinical situations. To assess whether the “ROCF tracing task” can be used to identify USN, patients with right brain damage were divided into two groups according to presence or absence of USN based on screening tests. A group of healthy subjects was also included. To examine characteristics of the ROCF tracing task, the tracing task and a copy task [21] were performed using ROCF (i.e. the “ROCF copy task”) and scores were compared among the three groups. In addition, correlations between scores on the tracing task and scores on conventional USN screening methods (the cancellation test, line bisection test, and ROCF copy task) were computed. Finally, to characterize performance of the ROCF tracing task in terms of sensitivity and specificity, a receiver operating characteristic (ROC) curve was computed for each test in the two groups of patients with right brain damage.

Subjects with right brain damage were recruited from patients admitted to Kirishima Rehabilitation Center of Kagoshima University Hospital. Healthy subjects were recruited from the general public.

The number of patients was set at 20 on the basis of cross-sectional studies of patients with right brain damage [22, 23].

The sample size of the healthy subject group was set at 40, double the number of right brain-damaged subjects, on the basis of cross-sectional studies of healthy controls and patients with right-brain damage [24, 25]. Thus, the patient and control sample sizes were different.

Inclusion criteria for patients with right brain damage were that they had had a first stroke, were right-handed, and were aged 20–80 years. Right brain damage was determined by consulting the medical history and by using medical information from the hospital where acute treatment was carried out, as well as by computed tomography images of the brain. Time since stroke onset was not considered in the patient inclusion criteria. Inclusion criterion for healthy subjects was being of age 20–80 years.

Exclusion criteria for both patients with right-brain damage and healthy subjects were a history of treatment for brain disorders (in the case of patients with right-brain-damage, a history of illness like stroke, brain injury, or surgery to the brain prior to the current stroke), mental illness or dementia which could affect the assessment results, or inability to understand the test description due to the effects of higher brain dysfunction or other factors.

We administered to participants various tasks described below, after obtaining written informed consent. This study complied with the Declaration of Helsinki and was approved by the Clinical Research Ethics Committee of Kagoshima University Hospital, Faculty of Medicine and Dentistry (No: 24–140).

Presence of USN was determined by impaired performance on the USN screening test. The USN screening test consists of the line cancellation test [26], star cancellation test [27], line bisection test [27], and copying a landscape constructed of five objects [15] this is the method of assessment used in our hospital [28]. Cutoff values for the line bisection test and star cancellation test were based on the Japanese version of the Behavioral Inattention Test. The line cancellation test was judged by two or more omissions on the left half of the sheet compared with the number of omissions on the right half. The copying test was judged by omission of at least one detail of the left part. Patients with right-brain damage who showed symptoms of USN in any of the tests were assigned to the “USN + group”; other patients with right-brain damage were assigned to the “USN − group”.

The ROCF tracing task was conducted immediately after the ROCF copying task. In the tracing task, subjects were presented with the ROCF sample and a sheet of carbon paper overlaid on a sheet of white paper, and instructed to trace the ROCF with chopsticks so as not to leave any tracing lines directly on the sample figure (Fig. 1a). At the same time, they were instructed not to trace the same spot more than once.

a ROCF tracing task; b Scoring respective 18 units of the ROCF and estimating the total, left (red), and right (blue) scores. Copying and tracing scores were estimated after the ROCF copying and tracing tasks were completed by the subjects

Both the tracing and copying tasks were evaluated conventionally with a maximum score of 36 points across 18 scoring units [29]. In addition, to identify differences between left and right structure copying and tracing, the left and right units were scored separately (Fig. 1b) (the left and right maximum scores were 12 and 16 points, respectively). In addition, the left–right ratio of correct responses was calculated as the Laterality Index (LI):

$$\mathrm\;(}) = \frac/12)}/16)}$$

For example, an LI below 1 indicates a greater proportion of correct responses in the right structures, and an LI above 1 indicates a greater proportion of correct responses in the left structures.

Further, to score the frequencies of overlapping (i.e., tracing the same item) in the respective copying and tracing tasks, each scoring unit of the ROCF structures was scored 1 point for a complete overlap and 0.5 points for an incomplete overlap, and was calculated as the total, left, and right “overlapping scores” by adding the number of times the unit was traced with an overlap.

Subjects were first interviewed about their education history and medical history related to the exclusion criteria, and tested using the Mini Mental State Examination-Japanese (MMSE-J) [30, 31]. The ROCF copying task and tracing task were then administered, on the same day.

The rules of the ROCF tracing task were explained to the subjects immediately after they had performed the ROCF copying task, and then they carried out the ROCF tracing task. If it was deemed that a subject did not understand the rules of the tracing task, the subject was asked to stop the task immediately and restart it after the rules were explained again. A subject was dismissed after having completed all of the tasks, and we then scored each task.

We compared right and left scores, LIs, and overlapping scores on the ROCF copying and tracing tasks among the three groups: healthy, USN − , and USN + . Primary outcome was the LIs on the ROCF tracing task, and the secondary outcome was the total overlapping score on the ROCF tracing task. The others were exploratory outcomes. The Shapiro–Wilk test was used to assess whether variables approximately followed a normal distribution; if lack of normality was judged, non-parametric tests (Kruskal–Wallis ANOVA followed by the Mann–Whitney U test with the Holm correction) were employed. A p-value of < 0.05 was considered to be statistically significant for the primary outcome. The effect size (r) was also calculated as r = √(t2/(t2 + df)).

To characterize the ROCF tracing task compared with the conventional USN screening test for subjects with right brain damage, we evaluated the correlations between scores or LI in the ROCF tracing task and those in the star cancellation test, line bisection test, or ROCF copying task, using Spearman's rank correlation coefficient. Furthermore, we performed Receiver Operating Characteristic (ROC) curve analysis to explore the performance, via the area under the ROC curve (AUC), of the scores and LI both in the ROCF tracing task as a USN detection test and in the conventional USN screening tests (i.e. the star cancellation test, line bisection test and the ROCF copying task). Tests with AUC greater than 0.9 were interpreted as highly accurate, 0.7–0.9 as moderately accurate, and 0.5–0.7 as rather inaccurate; an AUC of 0.5 represents a chance result [32].

All statistical analyses were performed using R (The R Foundation for Statistical Computing, version 4. 0. 2).