The sample included 178 older adults (28 men and 150 women), residing in the city of Palermo and its surrounding province, who voluntarily participated in this study. The inclusion criteria were designed to ensure that only older adults without known cognitive or neurological disorders took part in the study. In particular, the absence of cognitive and physical impairments was verified using a combination of self-reported medical history and indicators of functional independence in everyday life. Participants were recruited through the Provincial Health Authority of Palermo and were required to be autonomous in daily activities, to have no evident motor or sensory deficits that could interfere with testing, and not to be receiving treatment for psychiatric or neurodegenerative disorders. Individuals who reported previous diagnoses of dementia, Parkinson's disease, stroke, or other conditions known to affect cognition or mobility were excluded.

Sample had an average age of 70.66 years (SD = 6.14). (while education levels ranged from 4 to 23 years of education (M = 10.60, SD = 4.26). Weight ranged from 15.0 kg to 123.0 kg (M = 67.52; SD = 12.38), while height ranged from 1.40 m to 1.83 m (M = 1.61; SD = 0.07). Body Mass Index (BMI) ranged from 6.67 to 44.10 (M = 25.97; SD = 4.19).

The study was performed from November 2022 to January 2023 as part of an intervention evaluation for the Physical Activity Promotion & Domestic Accidents Prevention (PAP & DAP) Project, in collaboration with Palermo's Provincial Health Authority (ASP). This project, founded in 2021, is a multidisciplinary intervention that integrated a psychological, educational and physical activity approach for promoting health and well-being in over 60 population. This 3-year project provided theoretical and practical activities aimed at preventing the main risk factors for domestic accidents, promoting active lifestyles and improving quality of life. Specifically, training protocols were implemented to improve body balance, muscle flexibility, cardiorespiratory fitness, and muscle strength. The psychological and educational approach focused on awareness of the stereotypes and emotions associated with aging; the opportunities and resources of active aging; the perception of domestic risks and educational prevention strategies; and the motivation to adopt an active lifestyle. Recruitment took place from September to November 2022, using voluntary sign-ups via posters distributed by ASP representatives in different districts of Palermo. All participants provided written informed consent, participation was entirely voluntary, with the option to withdraw at any time. This study was reviewed and approved by the local Committee (124815-04/11/2022).

Participants completed a series of standardized tests and questionnaires designed to assess PA levels, cognitive functioning, emotional functioning, and psychological well-being. Paper-based questionnaires were self-administered, while digital versions were completed via a Google Form accessible on smartphones, under the supervision of the research team. The two formats (paper and digital versions) were identical in terms of item content, wording, and response scales. Mixed administration modes were adopted for practical reasons, particularly to facilitate the administration process. Motor assessments were conducted by a motor science specialist, whereas cognitive and emotional assessments were administered by a psychologist.

The International Physical Activity Questionnaire – IPAQ (47), Italian adaptation by Mannocci et al. (48), includes 7 items designed to measure the amount of PA performed in the previous seven days. Participants were asked to report the number of days and minutes they engaged in vigorous, moderate, and walking activities, as well as the time spent sedentary. These responses were used to calculate a score reflecting the Metabolic Equivalent of Task (MET) in minutes per week (MET min/week), which represents the energy expenditure associated with PA. One MET corresponds to the energy used while at rest. For each type of PA, a specific MET value is applied by multiplying the time spent on the activity (in minutes), the number of days, and a pre-assigned MET value. Walking is assigned a value of 3.3 METs, moderate activity 4 METs, and vigorous activity 8 METs. The final score is calculated by summing the METs for each type of activity (vigorous, moderate, or walking). Based on their total MET score, participants are categorized as inactive if the total is less than 700 METs, sufficiently active if between 700 and 2,519 METs, and active or very active if above 2,520 METs.

The Psychological Well-Being Scale – PWBS (49) was employed to assess participants' eudaimonic well-being. It was used the short Italian version of the scale (50) composed of 18 items scored on a 5-point Likert scale (1 = completely disagree; 5 = completely agree). A higher score indicates a higher level of well-being. The PWB scale is composed of six subscales, each evaluating a different dimension of well-being: self-acceptance, positive relationships with others, autonomy, environmental mastery, purpose in life, and personal growth. The Self-Acceptance subscale measures an individual's attitude toward themselves while the Positive Relations with Others subscale evaluates the quality of one's relationships. The Autonomy subscale gauges a person's capacity for making independent decisions, and the Environmental Mastery subscale assesses how well individuals manage and adapt to their surroundings. The Purpose in Life subscale measures a sense of having goals and meaning, while the Personal Growth subscale assesses continuous personal development.

The Clock Drawing Test – CDT (51, 52), Italian version by Caffarra et al. (53), is a cognitive assessment tool primarily used to screen for visuospatial and executive functioning deficits, as well as to evaluate general cognitive abilities. In the CDT, participants are asked to draw a clock, typically instructed to position the hands to a specific time (e.g., “10 past 11”) within a drawn circle, including the numbers in the correct order around the clock face. Scoring for the CDT considers the accuracy of number placement, the drawing of the clock hands, and the overall spatial organization of the clock face. Errors in any of these areas can indicate cognitive impairments associated with conditions such as dementia or other neurodegenerative disorders.

The Mini Mental State Examination – MMSE (54), Italian version by Measso et al. (55), is a cognitive screening composed of 30 questions that cover seven cognitive domains: temporal and spatial orientation, word registration, attention and calculation, word recall, language, and constructive praxis. Scores range from 0 to 30, with one point awarded for each correct response.

The Stroop test – ST (56), Italian version by Caffarra et al. (57), was administered to assess inhibitory self-control in the face of verbal interferences. The test has three sections: (1) participants read words that represent colors (e.g., blue, green, red); (2) they identify the color of dots (in blue, green, or red); and (3) they state the color of the ink (blue, green, or red) used in color words (e.g., blue, green, red). Completion times for each section were recorded, and two parameters were derived: Time (indicating the degree of cognitive interference, with longer times reflecting greater interference) and Errors (where a higher error count suggests greater difficulty in managing the interference).

The digit Span Task – Backward – DST-Bwd (58), Italian version by Monaco et al. (59), is a test used to measure working memory, or updating ability, which is the capacity to retain information while engaging in goal-directed cognitive tasks. In this test, participants are required to repeat a sequence of digits in reverse order, with the sequence length increasing progressively. The score is determined by the longest sequence (in terms of digits) accurately recalled.

The Digit Span Task – Forward – DST-Fwd (58), Italian version by Monaco et al. (59), is a test that assesses short-term memory. In this version, participants repeat a sequence of digits in forward order, with the sequence length gradually increasing. The score reflects the longest sequence of digits that the participant can correctly recall. This assessment is included in the cognitive testing to help screen out participants who may have potential short-term memory impairments.

The Meta-Emotional Beliefs Scale – CE (60), is a self-report that consists of 16 items rated on a five-point Likert scale, from 0 (not true) to 4 (definitely true). This scale examines individuals' beliefs regarding the perception, facilitation, understanding, and regulation of emotions. The CE score reflects the extent to which individuals believe that each aspect of emotion outlined in the EI ability-based model is significant and affects daily life. For example, it assesses beliefs such as whether sensations lead to emotions, emotions aid in thinking, emotions can blend together, or emotions can be regulated. The original version developed for adolescents was adapted for use with adults, and both exploratory and confirmatory factor analyses supported the four-branch structural model (χ²(164) = 22.2, p = 0.075, CFI = 0.99, TLI = 0.99, SRMR = 0.02, RMSEA = 0.03, 90% CI [0.00, 0.04]). Normative data were obtained from a sample of 963 adults (805 women and 158 men), and scores are standardized, with a mean of 100 and a standard deviation of 15.

The Emotional Self-Concept Scale – CME (60) is a self-report that assesses self-perceived abilities in perceiving, facilitating, understanding, and managing emotions. This scale includes 20 items, also rated on a five-point Likert scale from 0 (not true) to 4 (definitely true). The items ask respondents to assess their emotional abilities in everyday situations. The score represents how capable individuals believe they are in perceiving, using, understanding, and managing emotions in daily life. Initially developed and validated for use with pre-adolescents and adolescents, the instrument was subsequently adapted for adults. Exploratory and confirmatory factor analyses supported the four-branch structural model in the adult version (χ²(164) = 38.4, p < 0.001, CFI = 0.98, TLI = 0.97, SRMR = 0.02, RMSEA = 0.04, 90% CI [0.03, 0.06]). The normative dataset included 963 adults (805 women, 158 men), and scores are standardized, with a mean of 100 and a standard deviation of 15.

The Mayer-Salovey-Caruso Emotional Intelligence Test – MSCEIT (61), Italian validation by D'Amico and Curci (62), is an ability-based measure that evaluates emotional intelligence (EI) through maximum performance. The test assesses the four key branches of ability EI (63) and consists of 141 items organized into eight tasks, each pair corresponding to one of the four EI abilities: perceiving, facilitating, understanding, and managing emotions. The MSCEIT provides fifteen primary scores: eight subscale scores, four branch scores, two area scores (experiential and strategic), and an overall emotional intelligence quotient. Participants’ responses are scored using general consensus norms. Scores were calculated using the consensus-based scoring approach (61), which assigns higher scores to responses that align with the most frequently selected answers in the large normative sample. This approach reflects the assumption that emotional intelligence involves socially shared understandings of emotional situations, with correctness determined by the degree of agreement with the broader population. The normative sample used for consensus scoring consists of 3,185 adults (1,718 females and 1,487 males) and corresponds to the Italian standardization sample reported in the Italian validation of the MSCEIT (62). Scores are standardized, with a mean of 100 and a standard deviation of 15.

All statistical analyses were performed using IBM SPSS Statistics, version 25.0 (64). First, prior to the main analyses, assumptions underlying parametric tests were examined. Distributional properties of the variables were inspected, and assumptions of normality and homoscedasticity were evaluated. Second, descriptive statistics were calculated for all study variables and Pearson correlation coefficients were computed to examine the bivariate associations between PA levels, measured as Metabolic Equivalent of Task (MET), and psychological well-being, cognitive and emotional variables. Third, to derive a composite indicator of general cognitive functioning, a principal component analysis (PCA) was conducted on cognitive test scores. The PCA included measures of working memory, attentional and inhibitory control, and visuospatial abilities, while the MMSE was considered separately as a global screening measure. Cognitive variables were coded so that higher scores reflected better performance. Sampling adequacy was verified using the Kaiser–Meyer–Olkin measure and Bartlett's test of sphericity. A single-component solution was retained and component scores were saved for subsequent analyses examining associations with PA, emotional dimensions, and psychological well-being. Fourth, in addition to correlation analyses, participants were classified into three groups based on their level of PA (i.e., inactive, sufficiently active, and very active), according to the IPAQ scoring guidelines and total MET minutes per week. This grouping was used to examine whether different levels of PA were associated with differences in cognitive functioning, emotional dimensions, and psychological well-being. Accordingly, a series of one-way Analyses of Variance (ANOVAs) were conducted on the dependent variables related to cognitive performance (e.g., MMSE, Clock Drawing Test, Stroop Test, Digit Span Forward and Backward), emotional dimensions (e.g., Meta-Emotional Beliefs, Emotional Self-Concept, MSCEIT), and psychological well-being (e.g., the six Ryff subscales). Lastly, when statistically significant main effects were found, Bonferroni post-hoc tests were used to explore pairwise differences between groups. Effect sizes were reported using partial eta squared (partial η²).

Table 1 presents descriptive statistics (i.e., mean and standard deviation) and the Pearson correlation coefficients between the study variables. The correlation matrix highlights several relationships between PA, as measured by the Metabolic Equivalent of Task (MET), and various psychological, emotional and cognitive measures.

A significant positive correlation was found between MET and Digit Span Task – Backward (DST-Bwd: r = 0.30, p < 0.01), indicating that higher PA levels are associated with better working memory capacity. However, the association between MET and Digit Span Task – Forward was not significant (DST-Fwd: r = 0.14, p > 0.05). Moreover, no significant association between MET and the Clock Drawing Test was found (CDT: r = −0.10, p > 0.05).

MET also showed significant positive relationships with emotional and psychological measures. It was positively correlated with PWBS-Personal Growth (PWBS-PG: r = 0.20, p < 0.05), implying that individuals with higher PA levels tend to experience a greater sense of personal growth. Additionally, MET exhibited significant positive correlations with CE (Meta-Emotional Beliefs: r = 0.28, p < 0.05), suggesting that higher levels of PA are associated with more positive beliefs about emotions and emotional regulation. However, MET did not show significant correlations with other measures of psychological well-being or emotional dimensions.

To obtain a composite indicator of cognitive functioning, a principal component analysis (PCA) was conducted on cognitive test scores assessing working memory, attentional control, and visuospatial abilities (Digit Span Forward and Backward, Stroop Test indices, and Clock Drawing Test). The MMSE was excluded from the PCA and treated separately as a global screening measure. All cognitive variables were coded so that higher scores reflected better cognitive performance. Sampling adequacy was acceptable (Kaiser–Meyer–Olkin measure = 0.64), and Bartlett's test of sphericity was significant, χ²(10) = 48.0, p < 0.001. The PCA yielded a single component explaining 37.2% of the total variance (eigenvalue = 1.86). This component showed meaningful positive loadings across all included cognitive measures and was interpreted as an index of general cognitive functioning (GCF). The loadings of each cognitive measure on the first principal component are reported in Table 2.

Pearson correlation analyses were conducted to examine the associations between the general cognitive functioning index (GCF), PA levels, emotional dimensions, and psychological well-being. Correlation coefficients are presented in Table 3.

Results showed that GCF was positively associated with MET (r = 0.26, p < 0.05), indicating that higher levels of PA were related to better overall cognitive functioning. Moreover, significant positive associations emerged between GCF and the psychological well-being dimensions of personal growth (r = 0.37, p < 0.001) and autonomy (r = 0.30, p < 0.01), suggesting that better cognitive functioning was linked to greater perceptions of self-development and self-determination.

With respect to emotional dimensions, GCF showed a significant positive association with meta-emotional beliefs (CE: r = 0.32, p < 0.05). In contrast, no significant associations were observed between GCF and emotional self-concept (CME: r = 0.17, p > 0.05) or emotional abilities (MSCEIT: r = 0.32, p > 0.05).

Table 4 presents the results of one-way ANOVA tests comparing the psychological well-being and cognitive measures in inactive (25.3%), sufficiently active (32.9%) and very active (41.8%) groups. post hoc pairwise comparisons were conducted using Bonferroni correction to control for multiple comparisons.

Significant group differences were found in the Metabolic Equivalent of Task (MET) [F(2, 155) = 64.39, p < 0.001, partial η² = 0.45], with very active individuals (M = 5,497.29, SD = 3,937.70) showing significantly higher MET scores compared to sufficiently active (M = 1,371.39, SD = 503.98) and inactive individuals (M = 233.63, SD = 225.21). Additionally, significant differences were observed in Backward Digit Span Task (DST-Bwd) [F(2, 94) = 5.79, p = 0.004, partial η² = 0.11], with very active individuals (M = 3.85, SD = 1.20) outperforming sufficiently active individuals (M = 3.03, SD = 0.97).

Significant differences emerged for PWBS-Personal Growth [F(2, 147) = 3.33, p = 0.039, partial η² = 0.04], with post-hoc Bonferroni comparisons revealing that very active individuals (M = 11.84, SD = 1.88) reported significantly higher personal growth compared to inactive individuals (M = 10.92, SD = 1.91). No significant differences were found between groups on other PWBS subscales, including Positive Relations [F(2, 147) = 0.008, p = 0.924], Autonomy [F(2, 147) = 0.094, p = 0.394], Environmental Mastery [F(2, 147) = 1.30, p = 0.276], Purpose in Life [F(2, 147) = 1.22, p = 0.299], and Self-Acceptance [F(2, 147) = 0.076, p = 0.469].

No significant differences were found between PA groups for other cognitive measures, including MMSE [F(2, 38) = 0.84, p = 0.435], Clock Drawing Test [F(2, 95) = 1.93, p = 0.151], Stroop Test Time [F(2, 93) = 0.96, p = 0.386], Stroop Test Errors [F(2, 93) = 0.28, p = 0.756], Forward Digit Span Task (DST-Fwd) [F(2, 94) = 1.40, p = 0.251]. Additionally, no significant differences were found between levels of PA and Meta-Emotional Beliefs (CE) [F(2, 68) = 1.06, p = 0.351], Emotional Self-Concept (CME) [F(2, 68) = 0.82, p = 0.444], and Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT) [F(2, 38) = 2.94, p = 0.065].

While the findings of this study provide valuable insights into the relationship between PA, cognitive function, and emotional well-being in older adults, several limitations should be considered when interpreting the results. While significant correlations were observed between PA and cognitive/emotional outcomes, it remains unclear whether PA directly influences cognitive function and emotional well-being, or if other factors contribute to these relationships. Longitudinal studies are needed to establish causal pathways and better understand the long-term effects of PA on cognitive and emotional health. PA was assessed using IPAQ, which relies on self-reported data. Self-report measures are susceptible to biases such as overreporting or underreporting, as participants may inaccurately recall or exaggerate their activity levels. Objective measures of PA, such as accelerometers or pedometers, could provide more accurate data and strengthen the validity of the findings. The cognitive and emotional assessments used in this study may not fully capture all aspects of cognitive functioning and emotional well-being. Future studies could benefit from using a broader range of cognitive and emotional assessments to capture a more comprehensive view of these domains. It is also important to note that self-report measures were administered using both paper-and-pencil and digital formats. Although the two formats were identical in terms of item content, wording, and response scales, the use of mixed administration modes may represent a potential source of measurement variability. The study sample consisted predominantly of women which may limit the generalizability of the findings to men. Future studies should aim for a more balanced gender representation.