3. Results

3.1. Questionnaires
To save space Figure 2 shows only the distribution of the T-scores on the ADHD index as it is considered the most reliable index of ADHD. T-scores have a mean of 50 and standard deviation of 10; they are transformed from raw scores and used to compare the individual's answers to population norms [21]. The figure indicates that the index score varies along a continuum of severity and provides enough variance to test our interhemispheric interaction hypothesis using the dimensional approach. The T-score of 65 may be used as a clinical cut-off relative to the population. As can be seen, about 10% of the sample had ADHD index score of 65 (four participants) and higher (seven participants). In addition, answers on the questionnaire might be considered reliable because only 15 subjects had a score above seven on the inconsistency index of the CAARS purported to identify random or careless responding.
Like the CAARS, also the DASS is a quantitative measure with cut-off scores to characterize degree of severity relative to the population. The large majority of the sample scored within the normal range on severity (i.e., the categories normal plus mild) of mood symptoms; about 15% of the sample had high mood symptoms (see Table 2).
Pearson correlations between the overall ADHD symptomatology (ADHD index) and mood symptoms were r = 0.57, p < 0.001 for depression, r = 0.37, p < 0.001 for anxiety, and r = 0.50, p < 0.001 for stress indicating a moderate overlap of ADHD and mood symptomatologies.

3.2. Banich Letter Matching Tasks
Less than 5% errors were made; therefore, number of errors has not been taken into consideration. The reaction time analyses were run with and without the 15 subjects with inconsistent responding on the CAARS and run with and without the seven participants with mixed handedness. The outcomes did not essentially differ. Thus, data analyses on 112 subjects are presented below.
Pearson correlations were calculated between the overall mean RT (using the physical- and the name-identity task) and the ADHD index, inattention, hyperactivity, impulsivity, depression, anxiety, and stress. Only the ADHD index showed a positive trend with the overall mean RT (r = 0.17, p = 0.07). The finding indicates the higher the ADHD index score, the slower the RT performance. This was especially the case in the name-identity task (r = 0.20, p = 0.03) and more specifically for within hemisphere trials (r = 0.22, p = 0.01).
Because of the imbalance gender ratio in the sample, we have tested for gender differences. Neither the main effect of gender, on reaction time performance, nor its interactions with task nor trial type were significant (p ≥ 0.27). Therefore, we collapsed the mean RTs for males and females together (see Figure 3).
To investigate whether our reaction time data could be interpreted in terms of the Banich paradigm, a repeated measures analysis of variance was carried out on reaction time performance. The within subjects factors were task (physical-identity, name-identity) and trial type (within hemisphere, across hemisphere). The analysis revealed a significant main effect of task, F (1, 111) = 421.01, p < 0.001, η 2 p = 0.79, with faster RTs in the physical-identity task (M = 593) than the name-identity task (M = 743). According to the Banich paradigm, the interaction between trial type and task must be significant. This was the case in our data set: the interaction was F (1, 111) = 40.96, p < 0.001, η 2 p = 0.27; the RTs in the name-identity task were faster on across hemisphere trials (M = 721) compared to within hemisphere trials (M = 766), while RTs in the physical-identity task were equal for within hemisphere (M = 592) and across hemisphere trials (M = 593).
The left and right panels of Figure 4 present the distribution of interhemispheric interaction speed index (difference in mean RTs between within and across hemisphere trials) of, respectively, the physical- and the name-identity task. Both panels together show that the indices were normally distributed, statistically confirmed by the nonsignificant Shapiro-Wilk tests (W = 0.98, p = 0.08 for the physical-identity task and W = 0.99, p = 0.71 for the name-identity task).
The interhemispheric interaction speed indices for both tasks were used as dependent variables in two multivariable linear regression analyses. In the first, the ADHD index was the independent variable. The results indicated that the ADHD index was not a predictor (R 2 ≤ 0.03, p ≥ 0.36). In the second, the three key domains of ADHD (inattention, hyperactivity, and impulsivity) were the independent variables. The results revealed nonsignificant predictors (R 2 ≤ 0.04, p ≥ 0.28). Consequently, the ADHD symptomatology in isolation from mood symptomatology has no association with interhemispheric interaction. The picture becomes different when exploring the combination of both symptomatologies. Tables 3 and 4 present regression analyses including the aforementioned ADHD subscales with a backward selection of the DASS mood subscales. The resulting models were significant only in the name-identity task. The key domains of ADHD symptoms and stress explained about 10% of the variance of interhemispheric interaction index. Here the ADHD symptoms (especially hyperactivity) and stress were significant predictors of, respectively, fast and slow interhemispheric interaction.
In sum, on the basis of regression analyses it is concluded that only the combination of ADHD symptoms and stress is linked with the speed of interhemispheric interaction.
To explore in detail the effect of stress on ADHD symptoms and the speed of interhemispheric interaction, Pearson correlations were calculated between the ADHD index, hyperactivity scores, and the interhemispheric interaction index of the name-identity task in the group with high stress (score >14 on the stress subscale; n = 33) and in the group with low stress (score ≤14; n = 79) apart. In the group with high stress, faster interhemispheric interaction was correlated with higher scores on the ADHD index (r = 0.37, p = 0.03) and the hyperactivity subscale (r = 0.41, p = 0.02). In the group with low stress correlations were not significant with p values ≥0.39.
Finally, to test the ADHD-interhemispheric interaction link more thoroughly we compared the speed of the interhemispheric interaction of first and third tertile groups on the ADHD index, inattention, hyperactivity, and impulsivity subscales of the CAARS. Repeated measures analyses of variance with a within subject factor of task (physical-identity, name-identity) and a between subjects factor of group (low-score, high-score) indicated that group composition based on the ADHD index, inattention, and impulsivity subscales revealed no group differences. However, the high-score group on the hyperactivity subscale showed overall faster interhemispheric interaction than the low-score group; the main effect of group was significant F (1, 74) = 3.95, p < 0.05, η 2 p = 0.06. Post hoc analysis for the hyperactivity subscale indicated that the groups differed in the name-identity task, t (74) = −2.17, p < 0.03, but not in the physical-identity task; the mean interhemispheric interaction indices were 0.046 and 0.086 in the name-identity task, and they were −0.006 and 0.010 in the physical-identity task for, respectively, the low- and the high-score group.