Introduction

Executive functioning (EF) comprises a multifaceted set of frontally mediated, noninstrumental cognitive processes that control instrumental domains and behavior [1]. Executive disorders are thus frequently found in a variety of neurological conditions of different etiologies that affect cortical/subcortical frontal structures [2].

Although second-level specific psychometric tests are to be preferred when assessing EF [3], screening instruments for executive deficits are often useful in clinical settings, such as providing with an optimal trade-off between informativity and both sensitivity and a rapid administration [4].

The Frontal Assessment Battery (FAB) [5] is an EF screener that requires 5–10’ to administer and consists of 6 tasks assessing different EF facets: (1) concept formation and abstract reasoning (similarities); (2) mental flexibility (phonemic verbal fluency); (3) motor programming (Luria motor sequences); (4) sensitivity to interference (conflicting instructions); (5) inhibitory control (go-no-go); (6) environmental autonomy (prehension behavior) [5, 6].

The FAB is one of the most widely used EF screeners worldwide; its psychometric properties, clinical usability, and neural correlates have been thoroughly investigated [7].

In Italy, the FAB has been adapted and normed, as well as validated in both healthy and clinical populations [6, 8]. However, current FAB Italian normative data date back to more than 15 years ago and sociodemographic changes require norms updating [9]. Moreover, to the best of the authors’ knowledge, the FAB has been only validated against “non-executive” screeners in Italy [6]—e.g., the Mini-Mental State Examination (MMSE) [10]—whereas its association with “EF-loaded” screening measures, e.g., the Montreal Cognitive Assessment (MoCA) [11], has yet to be explored. In addition, FAB normative studies do not provide with norms for its subtests, despite this being an increasingly widespread approach for cognitive screeners in Italy [12, 13], as it allows greater flexibility for clinicians when using these instruments.

Accordingly, the two aims of the present study are (1) providing updated normative data for both FAB total and sub-test scales in a large Italian representative population sample and (2) validating the FAB and its subscales against the MoCA.

Subjects and methods

Participants

The sample consisted of N = 475 healthy Italian native speakers from different provinces of Northern Italy (Table 1). Participants had no history of neurological, psychiatric, or severe general medical conditions (i.e. severe internal and metabolic morbidities or systemic/organ failures). Studies that data come from were approved by the Research Evaluation Committee of the Department of Psychology of University of Milano-Bicocca on behalf of the Ethical Committee of the same university. Participants provided informed consent to participation and signed a data treatment disclaimer research purposes.

Table 1 Sample stratification for age, education, and sex (M/F: M=males, F=females)
Full size table

Materials

Global cognition was assessed via the MoCA [12, 14], which encompasses subtests evaluating EF (MoCA-EF), attention (MoCA-A), language (MoCA-L), memory (MoCA-M), visuospatial functions (MoCA-VS), and orientation (MoCA-O). Supplementary Table 1 provides the protocol for the current FAB. FAB items (N = 6) were grouped into 3 subscales: FAB-1 comprising the first two items (similarities and phonemic verbal fluency, linguistically mediated EF); FAB-2, comprising the second two-item set (Luria motor sequences and conflicting instructions, planning); FAB-3, comprising the last two items (go-no-go and prehension behavior, inhibition). All participants were administered the MoCA first and then the FAB.

Statistical analyses

According to previous normative studies [15, 16], the minimum sample size was set at N = 287 by means of a power analysis (α = 0.05; 1-β = 0.9; f2 = 0.05) for multiple linear regression (dfnumerator = 3) analyses [17] via the R 3.6.3 package pwr [18].

Normality assumptions on both background and cognitive raw variables were checked by evaluating skewness and kurtosis values (judged as abnormal if ≥|1| and |3|, respectively) [19].

Associations of interest between quantitative variables were assessed by means of either Pearson’s or Spearman’s techniques. When judged as relevant, Bonferroni correction for multiple comparisons was applied.

According to the equivalent scores (ES) method [20, 21], raw scores (RSs) were adjusted for significant intervening background predictors (or their transforms) via regression-based equations. The cutoff was identified by computing outer and inner tolerance limits (oTL; iTL). Adjusted scores (ASs) were then standardized into a 5-level quasi-continuous scale: ES = 0 (ASs ≤ oTL; “abnormal”); ES = 4 (ASs > Mdn; “normal”); ES = 1, 2, and 3 (oTL < ASs ≤ Mdn; respectively, “borderline,” “low-end normal,” “normal”).

Analyses were performed via SPSS 27 [22] and R 3.6.3 [23]. Regression studies, as well as computations of both TLs and ES thresholds, were implemented according to guidelines and software solutions described in [24].

Results

Background and cognitive scores are summarized in Table 2.

Table 2 Participants’ background and cognitive scores
Full size table

The vast majority of FAB scales were related to MoCA measures (Table 3): the strongest associations were found with MoCA-EF and total scores. FAB subtest scores were all associated with each other (0.21 ≤ rs(475) ≤ 0.24; p < 0.001), as well as with FAB total scores (FAB-1: rs(475) = 0.65; FAB-2: rs(475) = 0.63; FAB-3: rs(475) = 0.7).

Table 3 Correlations between FAB and MoCA scores
Full size table

Age was negatively related to FAB-1 (rs(475) =  − 0.3; p < 0.001), − 2 (rs(475) =  − 0.36; p < 0.001), − 3 (rs(475) =  − 0.3; p < 0.001) and total (rs(475) =  − 0.44; p < 0.001) scores, whereas a positive association with education was detected: FAB-1 (rs(475) = 0.35; p < 0.001), FAB-2 (rs(475) = 0.24; p < 0.001), FAB-3 (rs(475) = 0.31; p < 0.001), and total (rs(475) = 0.42; p < 0.001). No sex differences were found: FAB-1 (t(473) =  − 0.92; p = 0.357), FAB-2 (t(473) = 0.44; p = 0.66), FAB-3 (t(385.3) = 1.5; p = 0.25), and total (t(473) = 0.4; p = 0.689).

When simultaneously tested, age and education revealed to be predictive of FAB both total and subtest scores (age: − 0.17 ≤ β ≤  − 0.34; p ≤ 0.001; education: |0.15|≤ β ≤|0.33|; p ≤ 0.001). Cubic age and logarithmic education proved to be the most significant predictors of all FAB scales, with the exception of FAB-2 and FAB-3 that were best predicted by reciprocal education and quadratic age, respectively (Tables 4 and 5).

Table 4 Adjustment grid according to age and education for FAB total raw score
Full size table
Table 5 Adjustment grids according to age and education for FAB-1, FAB-2, and FAB-3 raw scores
Full size table

Selected correction factors and adjustment equations for FAB total and subtest RSs are displayed in Tables 4 and 5, respectively. TLs and ESs classifications for all FAB ASs are reported in Table 6.

Table 6 Equivalent Scores for the FAB-T, FAB-1, FAB-2, and FAB-3 adjusted scores
Full size table

Discussion

The present work provides Italian practitioners with updated normative data for the FAB and its subtests. These norms cover a wider age and education range and are drawn from a larger sample size (N = 475) than those of previous normative studies (N = 236 [8] and N = 364 [6]). Moreover, norms for FAB subscales are provided—this representing a previously unreported feature that supports an adaptive usage of the screener.

This study overall replicates previous findings with respect to the range of FAB scores’ predictors in the Italian population: the performance increases with higher educational attainment and decreases with aging, while no sex differences are detected[6, 8]. It has nonetheless to be noted that the cutoff reported here (12.03), despite being similar to that derived by Iavarone et al. [8] (11.54), is more conservative than Appollonio et al.’s [6] (13.5). This aspect might reflect sociodemographic changes that have occurred in the last two decades in northern Italian population.

To the best of the authors’ knowledge, this contribution is the first showing convergent validity between the FAB (and its sub-scales) and an “EF-loaded” cognitive screener—i.e., the MoCA—in Italian healthy individuals. The present results also support the notion of the MoCA being a screening instrument sensitive to EF deficits, when compared to other screeners, such as the MMSE [25].