The OCS was comprehensively designed for aphasia and neglect symptoms. This was achieved by adopting short high-frequency words, forced-choice testing procedures, vertical layouts, and multimodal presentations. It uses some tasks intended to assess more than one domain. For example, the “reading” task is used not only for memory but also for neglect symptoms. This feature was attributed to the short administration time of the OCS (within 15 minutes) to complete. It can also be administered from the acute phase of stroke, three days after onset. The original author, Nele Demeyere, specified minimum target number of subjects as 60 for statistically valid standards in the translation and linguistic validation process [26].

As for the criteria to determine impairment, the OCS uses 5th-percentile cutoffs. Mostly, the scores of 5th percentile represent impairment. However, according to the task, the scores of ≥95th percentile in the following three tasks also indicate impairment: (1) “space asymmetry,” (2) “object asymmetry,” and (3) “executive function total” scores. In the “space and object asymmetry” tasks, which detect neglect, high positive and negative scores represent left and right spatial neglect, respectively. Thus, both the 5th percentile and 95th percentile could be used as criteria for the norm in these tasks.

In “executive function” tasks, which have four sorts of scores, stimuli are composed of seven circles and triangles each, which might distract examinees. One point is assigned to each correct connection. The next point is given if the subject correctly completes the connection on the next try, even for existing errors at some points in the previous line connection drawing. Therefore, the maximum number of points was six each for circle and triangle connections. Moreover, the alternative circle and triangle connecting task, the “executive (mixed)” task (maximum of 13 points) is to be followed. Then, the “executive function total” score can be obtained by subtracting the “executive (mixed)” score from the sum of scores in the “executive (circles)” and “executive (triangles)” tasks. The higher the score of the task, the greater the loss of executive function; thus, the score of the 95th percentile is the cutoff score.

Two licensed clinical psychologists in their psychology PhD program, who were proficient in English, translated the original OCS [15], and another clinical psychology specialist who received a PhD in the United States conducted a reverse translation without reading the original version. The original writer, Nele Demeyere at Oxford University, assisted with the entire translation process and provided the final approval.

The basic structure of the test and the scoring rules of the original version were applied to the K-OCS without any changes. However, as this study aimed to adapt the OCS to Korean culture and language characteristics, some objects and words were replaced. A pilot study was conducted with 15 adult participants without disabilities, and a few nationality-adapted modifications were made. The first modification was required for the “picture naming” task in the language domain. The objects provided for the naming test should be challenging enough for patients with cognitive impairments but, simultaneously, familiar enough to be recognized. Some of the objects in the original version, such as “pear” and “filing cabinet,” were reported to be unfamiliar to most of the participants, especially to people older than 40 years who were expected to be the main subjects of this test. To select valid figure objects as options, we performed a validation study on the major candidate age population. Correct answer rates for the five candidate objects in the pilot test, “hippopotamus,” “watermelon,” “axe,” “pomegranate,” and “fire extinguisher” were 87.5%, 100%, 100%, 72.4%, and 87.5%, respectively (n=15). As the test is supposed to measure one’s naming function in usual daily living, rather than knowledge level, the word with accuracy under 80%, that is, “pomegranate” was excluded. The final selected number of the items was four—“hippopotamus,” “watermelon,” “axe,” and “fire extinguisher”—similar to the original version.

The second modification was made for the “reading” task of the language domain and the “recall and recognition” task in the memory domain, both of which used the same sentence. The principles of the sentence construction proposed by the original author were as follows: (a) placing the high neighborhood words (i.e., words that can change meaning by changing or deleting one letter, such as “cat” and “pat” in English) leftmost in four lines of the sentence; (b) the sentence consisting of 14 words including particles, and 4 of those should be infrequent words; and (c) phonically irregular words, such as “islands,” “quay,” “colonel,” and “yacht” should be included. To generate a new Korean sentence based on these principles, several adjustments and agreements with the original author were required to address the phonetic and syntactic differences between Korean and English. First, as there are few phonetically irregular words in Korean, infrequently used words were adopted instead of irregular words based on the frequency of modern Korean word usage [27]. The second adjustment was to make sentences in Korean similar to the original sentence in speech duration rather than word count itself, owing to mismatches in the syntax between the two languages. To match the speech duration, we added one more syllable for the pilot trial. Thus, 15 syllabi were included in the sentence, while the standard number of syllables in the original English sentence was 14. In a preliminary study (n=15), when the number of words in a sentence was 14, the mean value of the “sentence reading” performance was 13.97. When the number of included words was 15, its mean value was 14.65, indicating that an increase in the number of words in a sentence exerted no significant effect on the results of the “recall and recognition” tasks. Accordingly, a 15-syllable sentence is created.

Both the OCS and K-OCS include the following materials for testing: (1) a test booklet for multiple-choice items and examples, (2) a patient pack of paper-and-pencil task worksheets, (3) an easy scoring template, and (4) a user manual. The patient pack contains a visual snapshot report (Fig. 1), where impaired domains are marked by coloring in the blank spaces of the task for abnormal scores. Furthermore, detailed information about the patient can be shared by leaving comments adjacent to the figure, such as comments about the patient’s mood, physical state, or attitude that deserve consideration. The normative data and additional materials of the K-OCS can be downloaded from https://process.innovation.ox.ac.uk/clinical/p/ocs/questionnaire/1.

The study protocol was approved by the Institutional Review Board of Duksung Women’s University (no. 2020-007-015-A) in Seoul and Yongin Sevrance Hospital, and the CHA Bundang Medical Center (no. 2020-07-0202) in Gyeonggi Province. Written informed consent was obtained from all participants in the study. Participants without known disabilities were recruited from September 2020 to March 2022 via notice board announcements at Duksung Women’s University, Yongin Sevrance Hospital, and CHA Bundang Medical. A total of 97 volunteers who were more than 30 years old, and lived in Seoul or Gyeonggi Province, participated in this study. All participants were personally interviewed and screened using the Christensen Health Screening criteria [28], Korean version of MoCA (K-MoCA) [29], and the Beck Depression Inventory the 2nd version (BDI-II) [30]. Patients with neurological diseases, brain damage, visual and speech impairments, or psychiatric history were excluded. The K-OCS was administered to those who were finally enrolled after screening for eligibility. When the K-MoCA and K-OCS were conducted on the same day, the time interval was set to at least 30 minutes.

For data analyses, IBM SPSS Statistic 21, R4.1.3 version (IBM Corp.) was used. Descriptive statistics were used to demonstrate the demographic features of the study population, as well as the mean, median, standard deviation, and range of each K-OCS task score. To verify the goodness of fit, Kolmogorov–Smirnov and Shapiro–Wilk normality tests were conducted. Correlations between K-OCS scores and demographic variables, such as gender, age, and education level, were analyzed using Spearman’s correlation (one-sided test) due to the distribution of K-OCS values that deviated from normality. Finally, we computed the percentile values of each task to obtain the cutoff scores by performing a frequency analysis [25]. In addition, we performed logistic regression analysis to verify whether the participants’ age and education level could predict the results of each task.

Among 135 voluntary participants, 8 scored below the normal cutoff value of the K-MoCA according to age [29], and 15 participants were determined to have depression with BDI-II>18 points (i.e., the cutoff score) and were excluded from the final analyses. Therefore, the data from 97 participants (33 male [34.0%] and 64 female [66.0%]) were included in the final analysis. Among them, 15 were enrolled in both the pilot and main studies. The age ranged from 35 to 74 with a mean age and standard deviation of 54.3±9.7. Their education level ranged from 5 to 18 years, and the mean education level and standard deviation were 13.8±2.7 years. Demographic characteristics and the results of MoCA and BDI-II by age group are shown in Table 2. Significant differences were observed in the MoCA scores between the <50- and >59-year age groups (p=0.001), whereas no differences were found in education level or BDI-II scores.

According to interim analysis after assessing 30 cases, the score of the “reading” task was found to be lower than that of the original English version [15] and other reports [18,21]. Thus, we edited the sentence, which was the main stimulus of the task, to increase its score level in the middle of the study. Consequently, the data from 30 participants, who were examined before the sentence amendment were excluded, and those from the other 67 participants were included for analyses of “reading” and verbal memory “recall and recognition” tasks where the same revised sentence was used. The mean scores for each age and years of education group for the K-OCS task are presented in Table 3. Kolmogorov–Smirnov and Shapiro–Wilk normality tests showed that the tasks, except for “semantics” and “executive task (triangles),” did not meet the normality criteria (p<0.001). Finally, the score distribution and cutoff value with criteria of the 5th or 95th percentile for each task were obtained (Table 4).

The cutoff scores of previous studies conducted in Denmark, the UK, and Italy are presented in Table 5. Most task scores in the current study were comparable to those of large-scale studies conducted in other countries. Tasks that showed different cutoff scores with ≥1 point from all of the other countries were “sentence reading” and “imitation.”

We conducted a logistic regression analysis to verify whether years of education could predict the K-OCS scores. Cutoff scores of “semantics” and “executive task (circles)” had the maximum values, and the cutoff score of “recall” had the minimum value, of the score ranges; thus, these were excluded from the analysis. According to the regression model, all the tasks, except for the “imitation” task (odds ratio>upper confidence interval, p=0.05), did not predict task scores regarding education level and age.