Improvement in cognitive performance and mood in healthy older adults: a multimodal approach

Abstract

The aim of this research was to evaluate if being involved in a programme that integrates physical, mental, and social activities could help to reduce the impacts of cerebral ageing on cognitive functions. Fifty healthy adults over 65 years of age and without cognitive impairment were randomly assigned to either an experimental or a control group; subjects were equally divided by age, sex, schooling, physical health, mood, and social integration. For 6 months, the experimental group had biweekly meetings, participating in a multimodal approach based on a combination of simultaneous physical, mental, and social activities. At pre-test and post-test, both experimental and control subjects underwent a neuropsychological assessment, including tests to measure attention, verbal and spatial memory, language, constructional praxis, executive functions, processing speed, and intelligence. Furthermore, a visual analogue scale was used to examine well-being and mood states. Compared to the pre-test levels and the control subjects, the performance of subjects in the experimental group significantly improved in several neuropsychological tests, including attention, processing speed, memory, and executive functions, as well as mood state. Even in older subjects without cognitive impairments, a multimodal approach based on simultaneous physical, mental, and social activity can be a useful strategy that has beneficial effects on mood and cognition. The results suggest that an active lifestyle may protect against cognitive decline in ageing.

Introduction

It is commonly believed that cerebral ageing is associated with a decline in cognitive functioning. However, the increase in overall lifespan has allowed us to observe a noteworthy inter-individual variability in ageing (Reuter-Lorenz 2002). Typical ageing manifests itself as a reduction in cognitive ability, especially in memory, processing speed, and executive functions, but does not imply a loss of independence in daily life (Park and Bischof 2013). However, ageing characterized by progressive cognitive and behavioural decline is likely due to a pathological process instead, such as Alzheimer’s disease (Giannakopoulos et al. 2009). However, there is also the possibility of so-called successful ageing, characterized by the maintenance of adequate cognitive functioning despite the passing of time (Rowe and Kahn 2015); this type of ageing correlates with a functional brain reorganization (Cabeza 2002; Eyler et al. 2011). In addition, some individuals, who have been called ‘superagers’, show ‘youthful’ memory abilities (Gefen et al. 2014). Initial investigations have shown that they have distinct genetic, anatomic, and histopathologic markers and that their brains exhibit fewer markers of Alzheimer’s disease (Rogalski et al. 2013). An important theoretical advance was the finding that subjects exhibiting successful ageing may present neuropathological anomalies comparable to those found in subjects who suffer from dementia (Snowdon 1997). To justify this finding, a reserve capacity concept, which is defined as the capacity of mitigating the effects of ageing or disease, has been introduced (Stern 2012): a brain reserve would be available in relation to the structural characteristics of the individual’s nervous system (Valenzuela and Sachdev 2005), while a cognitive reserve would be related to the higher functional capacity of the individual’s neural networks (Tucker and Stern 2011). The Scaffolding Theory of Aging and Cognition (STAC) hypothesizes that with the increase in age, a balance occurs between damaging events and those of repair and compensation (Park and Reuter-Lorenz 2009). This hypothesis becomes plausible when examining the plasticity of the nervous system (Barulli and Stern 2013). Indeed, neuroscience research has documented that the nervous system is continuously ready to reorganize itself in response to stimuli coming from its environment, and thus, the organization of the nervous system depends in large part on the experiences which have taken place over the course of a lifetime (Maguire et al. 1997). For example, it is known that musical training is a powerful means of reorganization in brain structures (Bangert et al. 2001). Playing musical instruments is associated with significant changes in grey matter, as well as with the increased development of critically important neural fibre bundles, such as the arcuate fasciculus and the corpus callosum (Wan and Schlaug 2010). Clinical observations and neurophysiological and neuroimaging studies converge in disclosing that musicians and non-musicians possess differing structures of neural connections in relation to different levels of musical competence (Chang 2014). Therefore, thanks to neuroplasticity, training can strengthen the efficiency of synapses and modify the interactions between neural networks (Bryck and Fisher 2012). As a result, brain function can become more efficient at counterbalancing the consequences of disease or the effects of ageing (Boyke et al. 2008).

Accordingly, identifying variables that may enhance the cognitive reserve has become a fundamental goal in research. The aim of these investigations has shifted from classic studies of the risk factors (alcohol, neurotoxic factors, etc.), which should be avoided, to the identification of ‘neuroprotective’ strategies, which can favourably direct plasticity to cope with anatomical and functional age-related changes (Mejía et al. 2017; Yaffe et al. 2009). The factors considered crucial in preserving or improving cognition can be put into three important categories: physical exercise, cognitive training, and social activities (Kolb et al. 2010; Tucker and Stern 2011).

The effects of physical exercise have been extensively studied (Colcombe and Kramer 2003), and aerobic exercise is known to positively affect cognitive processes such as executive functions, attention, psychomotor speed, memory, and visuo-spatial ability (Angevaren et al. 2008). Aerobic fitness may be assessed in a variety of ways, including walking as a preferred and recommended physical activity for middle-aged and older adults (Farren et al. 2015). A number of studies have tried to determine the relationship between the frequency, intensity, and duration of walking programmes and cognitive functions (Soumaré et al. 2009) and suggest that even a routine physical task (e.g. outdoor walking at a usual pace for at least 1.5 h/wk) may be sufficient for improving cognitive performances both in healthy older subjects (Gomes-Osman et al. 2018; Weuve et al. 2004) and in adults with dementia (Rolland et al. 2007; Kemoun et al. 2010). There are many physiological mechanisms that could justify the neuroprotective effects of physical activity. For instance, an improvement in cognitive functioning was in fact associated with increased cerebral circulation, structural changes of grey and white matter, and increased connections between regions sensitive to ageing (Colcombe et al. 2006); furthermore, other studies have documented an increase in the levels of neurotrophins and a reduction in abnormal protein deposits and factors involved in inflammation (Phillips et al. 2014). More generally, the processes involved in angiogenesis and neurogenesis seem to be more efficient on cognitive functioning in terms of physical activity (Stillman et al. 2016).

Regarding cognitive training, an equally substantial amount of research has documented specific positive effects of trained cognitive functions (Ball et al. 2002; Park and Bischof 2013; Schmiedek et al. 2010). For example, the Advanced Cognitive Training for Independent and Vital Elderly Overall (ACTIVE) study, using three types of training—for memory (verbal episodic memory), reasoning (ability to solve problems that follow a serial pattern), and speed of processing (visual search and identification)—demonstrated that cognitive interventions helped healthy elderly individuals to perform better on multiple measures of the specific cognitive ability for which they were trained (Rebok et al. 2014). This strong clinical evidence regarding the impacts of cognitive training on the ageing brain is confirmed by a number of neuroimaging studies, showing that improvements in memory, attention, processing speed, executive functions, and fluid intelligence tend to correlate with anatomo-functional reorganization (Cabeza et al. 2002; Clark et al. 2017).

The combination of physical and cognitive training is considered to be equally efficient, but the data on the hypothesized superiority of this combination as compared to a single type of training are contradictory (Rahe et al. 2015). It has been suggested that an increase in cerebral blood flow from physical activity can strengthen the effects of additional mental activity (Oswald et al. 2006). The effects of the two types of training should be synergistic and combine the supposedly favourable actions of physical activity on neurogenesis and angiogenesis with the effects of cognitive activity on synaptogenesis (Theill et al. 2013). In addition, this modality is in fact a dual-task procedure and therefore may stimulate the functional integration of those brain regions solely involved in the control of the two activities (Eggenberger et al. 2015). However, there is still not a sufficient amount of literature to establish the difference between single and double stimulations and between simultaneous and sequential combinations of the two kinds of training (Shatil 2013).

The data in favour of the positive effects of physical and cognitive training on brain health and on the avoidance of cognitive decline and dementia seem sufficiently convincing, but remain open to query (Willis and Schaie 2009; Young et al. 2015). One of the issues is represented by the different research methodologies: physical training can be aerobic or anaerobic, while cognitive training may concern extremely different functions. In addition, combined physical and cognitive training can be done simultaneously or sequentially, and the training can also vary in regard to duration and intensity, if it is done individually or in a group, etc. Objectively, it becomes difficult to untangle the numerous variables that may affect the study results.

Furthermore, recently researchers have invited the use of activities which can be inserted naturally into the daily life of individuals and thus forego the characteristics of purely formal exercise (Verghese et al. 2003). From this point of view, the methodological differences between cognitive training and mental stimulation have been stressed: the first aims to develop specific cognitive abilities (for example, working memory), while the second instead includes all situations that non-specifically promote mental activity, such as participation in leisure activities (Kelly et al. 2014). Leisure activities may be defined as something that an individual chooses purely for pleasure and well-being to escape from the commitments of work or daily life (Kuykendall et al. 2015). Examples of this could be learning an instrument, chess, literature and writing, crossword puzzles, card or board games, participating in a study group or volunteering (Lustig et al. 2009). This approach, proposed recently enough to be classified as ‘non-traditional’ (Park et al. 2007), has been defined as ‘multimodal’ because it involves the integration of different types of interventions in order to implement an active lifestyle (Lustig et al. 2009). A particular emphasis is placed on group activities aimed at acquiring new skills. In other words, the two crucial components that should characterize a multimodal approach seem to be learning something new and social interaction. In regard to the first factor (mental stimulation instead of cognitive training), improved cognitive performance was shown in subjects involved in ‘productive engagement’ tasks (for example, attending courses in photography, bridge, computers, or a second language) compared to subjects involved in ‘receptive engagement’ tasks (such as activities that they were already capable of and did not require them to acquire any new knowledge or skills) (Stine-Morrow et al. 2007). For the second component (group rather than individual activities), the favourable role of interpersonal relationships in ageing is by now widely recognized, even if the underlying mechanisms are not entirely clear (Bennett et al. 2006). Conversely, social isolation and feeling lonely, although independent from each other, are two factors which have negative implications on health, psychophysical well-being, and lifespan (Holt-Lunstad et al. 2015). Social network variables (including network size, frequency of social contacts, satisfaction with social contacts, and social support) have been significantly associated with cognitive status and with odds of cognitive impairment (Fankhauser et al. 2017). From this point of view, several investigations suggest the importance of social activities such as organizing a theatre performance (Noice et al. 2004). In the Experience Corps Project, subjects provided assistance in an elementary school, both as organizational support and aiding in students’ literacy (Carlson et al. 2009). In the Odyssey of the Mind Project, participants regularly carried out group problem-solving activities for several months (Stine-Morrow et al. 2007). In the Synapse Project, older adults participated in a group requiring new tasks, such as learning to quilt or digital photography or both (Park et al. 2014).

In these studies using a multimodal approach, the improvements primarily concerned memory, problem-solving, and executive functions and appear to be correlated with changes in cerebral activation (Carlson et al. 2009). However, not all of the meta-analyses agree with the hypothesis that training can protect against cognitive decline in ageing, and an objective difficulty exists in comparing the different methodologies used in the various experiments (Bamidis et al. 2014; Young et al. 2015). Moreover, recent research such as the FINGER study demonstrated a relationship between a multi-domain lifestyle intervention and cognitive benefits (Kivipelto et al. 2017).

In the present research, efforts were made to verify if a multimodal approach based on the combination of simultaneous physical exercise, mental stimulation, and social activity could have a positive influence on the cognitive functioning of older subjects without cognitive deterioration. The working hypothesis predicts that, in contrast to the control subjects, trained subjects will show a significant improvement in cognitive performance compared to the pre-test assessment.

Method

Subjects

The participants in the study were recruited through a notice distributed at senior fitness clubs and CVAs (senior recreation centres) in the province of Perugia. The invitation was extended to subjects ≥ 65 interested in participating in research to monitor their cognitive ability. The 72 people who responded were preliminarily examined to evaluate their cognitive functioning, using the Mini Mental State Examination (MMSE), as well as their general health, using clinical history, the subjects’ personal opinions about their state of health, and available clinical reports.

The inclusion criteria were as follows: aged ≥ 65 (at the beginning of the study), a score of ≥ 24 on the MMSE, regular physical activity (any kind of physical activity at least once a week), independence in daily life (judged on the basis of an ad hoc clinical interview), and a sufficient amount of social participation (any kind of social activity at least once a month). The exclusion criteria were the presence of cerebrovascular, cardiovascular, metabolic, oncologic or neuropsychiatric diseases, or the use of drugs which could influence performance, such as antidepressants, anxiolytics, or thyroid treatments (oral antihypertensive and diabetic treatments sufficient for keeping clinical conditions under control were not considered criteria for exclusion).

The 50 subjects enrolled were randomly assigned to one of the two groups: 25 in the experimental group and 25 in the control group. To keep the asymmetry of the groups to a minimum, the subjects were preliminarily divided into three age brackets: 60–69, 70–79, and 80–90. Exceptions to the randomization were the presence of more than one member of the same family (for example, husbands and wives were always put in the same group) and any obstacles related to necessary attendance for the purpose of the experiment (for example, subjects who could not guarantee the necessary regularity of activity were always put in the control group). All of the participants gave informed consent and agreed to undergo two neuropsychological evaluations 6 months apart.

The demographics of the two groups are illustrated in Table 1.

Table 1.

Demographics of study participants

	Number	Sex	Age	Education	MMSE
Trained subjects	25	11 M 14 F	72.88 (± 4.78)	12.08 (± 3.86)	26.35 (± 1.33)
Control subjects	25	12 M 13 F	72.32 (± 4.18)	12.64 (± 4.33)	26.52 (± 0.93)

MMSE = age and education-corrected Mini Mental State Examination. Mean and standard deviations of age, education, and MMSE

M male, F female

Procedure

The treatment of the experimental group required two 60-min sessions a week for a period of 6 months. During each session, physical activity associated with mental training was done using the ‘Walk and Learn’ method (https://walexperience.com). This method includes aerobic activity consisting of walking at a steady pace while respecting an individualized rhythm based on age, weight and routine physical exertion. The walk is taken wearing specially designed socks on a soft mat, which is made with special patented characteristics (natural materials, such as a base of rubber to guarantee perfect stability, and on the surface, a soft material with a high percentage of cotton that is easily sterilized). The track is set up inside a gym, lit by the natural light streaming through the many windows, and in an elliptical shape on which the subjects repeatedly walk around. Throughout the walk, the subjects wear headphones and listen to audio tracks, which have been researched and created purposely to meet the most varied interests. The headphones are wireless and the latest generation, guaranteeing a high-quality listening experience, perfectly isolated from the outside environment while maintaining optimal lightness and comfort. The tracks are read by professional actors and are interrupted by brief musical intervals. The catalogue (‘Culture in Movement’) offers more than 200 titles with varying topics that include art, science, fashion, sport, etc. The choice of text to listen to was discussed in the group, and the possibility for individual choice was given based on personal preferences. At the end of the sessions, the subjects participated in a discussion, led by a psychologist, to share their experience; this was set up so that anyone could express his/her opinion or talk about anything new they had learned about the subject that they chose. Another purposely planned opportunity for the subjects to socialize was that they were brought to and taken home from the sessions together by bus (Crespi 2014).

The subjects did not receive any monetary compensation and were they asked to spend any money for this research (e.g. the transport was provided by the organization). Everyone (including the subjects in the control group) was informed of the findings from the neuropsychological examinations at the end of the study.

Neuropsychological assessment

At pre-test and post-test, subjects underwent a neuropsychological evaluation, including a standardized battery of tests assessing general cognition, language, praxis, attention, verbal and non-verbal memory, executive functions, and fluid intelligence. Table 2 presents the tests used in the assessment battery along with the cognitive functions that are assessed by each test. Language examinations were carried out, especially considering the fact that tests are strongly influenced by the impairment of executive functions, such as fluency tests and resolution of lexical ambiguity (Bilenko et al. 2009; Lezak 1995). In addition, a visual analogue scale was used to examine subjective health and mood states (Lezak 1995; Luria 1975).

Table 2.

Neuropsychological measures and the cognitive functions they assess

Test	Abbreviation	Functions examined	References
Mini mental state examination	MMSE	Orientation, concentration, language, praxis, memory	Measso et al. (1993)
Raven’s coloured progressive matrices test	CPM	Culture fair test of non-verbal fluid intelligence	Basso et al. (1987)
Digit symbol substitution test	DSST	Processing speed, attention, concentration	Rosano et al. (2016)
Digit span test	DST	Memory span	Monaco et al. (2013)
Digit span backwards task	DSBT	Working memory	Monaco et al. (2013)
Cancellation test	CT	Sustained attention, processing speed	Spinnler and Tognoni (1987)
Stroop colour–word test	SCWT	Divided attention	Scarpina and Tagini (2017)
Paired associate learning test	PALT	Verbal learning	Zappalà et al. (1995)
Corsi block-tapping task	CBTT	Visuo-spatial memory span	Monaco et al. (2013)
Rey–Osterrieth complex figure B	ROCF	a. Copy	Luzzi et al. (2011)
		b. Immediate visuo-spatial memory
		c. Long-term visuo-spatial memory
Controlled oral word association	COWA	Verbal phonemic fluency	Lezak (1995), Zappalà et al. (1995)
Category fluency test	CFT	Verbal semantic fluency	Lezak (1995)
Humpty dumpty test	HD	Ability to resolve homonymous ambiguity	Piccirilli et al. (2015)

Statistical analysis

The statistical analysis was performed using the SPSS 17 software package (SPSS, Inc., Chicago, IL). The normality of the data distribution was analysed with the skewness and kurtosis index. To align the data, a logarithmic transformation of the scores was applied. Pre-test patients’ performances on cognitive measures were assessed using the paired two-sample t test. To study the effects of the intervention, an independent two-sample t-test (experimental vs. control group) and a repeated measures ANOVA were conducted with group as a between factor and time (pre-test and post-test) as a within factor.

The difference between pre-test and post-test was calculated for the experimental and control groups by means of Cohen’s d effect size.

To further investigate the differences between pairs of observations (pre-test and post-test), each subject completed the sign test. The sign test can be used to explore paired data. Considering the paired differences, under the null hypothesis, the mean of the differences between the pre-test and the post-test would be zero (Whitley and Ball 2002).

Results

First, as expected from the inclusion criteria, there were no differences between the experimental and control groups for sex, age, education, perceived state of health and mood, or amount of social interaction. Moreover, the level of cognitive performance on the pre-test was the same in the two groups, with the sole exception of copy of Rey–Osterrieth complex figure B, in which the experimental group showed significantly higher scores (t = 2.145; df = 48; p < 0.05).

Table 3 shows the results of the statistical analyses. The differences in test scores within samples were detected using the paired two-sample t-test.

Table 3.

Differences in the outcome between groups (A = trained subjects; B = control subjects) using the t test and repeated measure ANOVA (p values are referred to logarithmic transformation dataset)

Test	Group	Pre-test		Post-test		T test		T test effect size	Repeated Measures ANOVA F (df1 = 1, df2 = 1)
		Mean	SD	Mean	SD	t (df = 24)	p-value	Cohen’s d
MMSE	A	26.35	1.32	26.98	1.15	− 2.095	0.047*	− 0.4	n.s.
	B	26.52	0.93	26.76	1.37	− 1.008	0.323	− 0.2
MPC	A	31.32	3.76	32.46	3.22	− 2.716	0.012*	− 0.5	n.s.
	B	32.58	2.78	32.54	3.64	0.22	0.827	0.0
DSST	A	24.76	8.62	28.28	7.43	− 2.589	0.016*	− 0.5	T
	B	25.00	7.96	25.56	7.49	− 0.589	0.561	− 0.1	F = 5.655 p = 0.021*
DST	A	5.51	0.71	5.58	0.73	− 0.469	0.643	− 0.1	n.s.
	B	4.49	0.60	5.20	0.88	2.104	0.046*	0.4
DSBT	A	4.42	0.99	4.36	0.85	0.202	0.842	0.0	n.s.
	B	4.21	1.02	4.51	0.79	0.086	0.932	0.0
CT	A	53.74	5.48	56.25	4.74	− 2.760	0.011*	− 0.6	T * G
	B	54.11	4.11	53.59	5.02	0.617	0.543	0.1	F = 5.466 p = 0.024*
SCWT	A	41.00	19.02	40.40	13.23	− 0.609	0.548	− 0.1	n.s.
	B	44.08	16.62	46.24	19.23	− 0.502	0.620	− 0.1
PALT	A	9.14	3.14	10.40	3.21	− 1.773	0.089	− 0.4	n.s.
	B	9.18	3.31	9.28	3.14	− 0.223	0.82	0.0
CBTT	A	4.14	0.75	4.54	0.57	− 3.970	0.001***	− 0.8	T * G
	B	4.51	0.79	4.39	0.96	1.047	0.306	0.2	F = 10.057 p = 0.003**
ROCF.a	A	30.19	1.15	30.16	1.46	0.110	0.914	0.0	G
	B	29.25	1.88	28.97	1.91	0.738	0.468	0.1	F = 8.207 p = 0.006**
ROCF.b	A	22.88	4.17	24.74	4.55	− 1.920	0.067	− 0.4	n.s.
	B	22.19	3.58	22.42	5.05	0.115	0.909	0.0
ROCF.c	A	21.99	4.37	24.37	5.56	− 2.058	0.051*	− 0.4	T * G
	B	21.58	4.33	21.34	4.98	0.520	0.608	0.1	F = 3.765 p = 0.058
COWA	A	36.68	8.45	40.20	9.33	− 3.564	0.002**	− 0.7	T
	B	36.08	8.19	39.96	10.76	− 1.325	0.198	− 0.3	F = 7.951 p = 0.007**
CFT	A	43.80	9.01	46.88	8.28	2.244	0.034*	− 0.4	T
	B	43.16	7.01	44.60	6.42	− 1.406	0.173	− 0.3	F = 6.946 p = 0.011*
HD	A	6.04	1.30	6.64	0.70	− 2.506	0.019*	− 0.5	T
	B	6.36	1.07	6.50	1.06	− 0.806	0.428	− 0.2	F = 6.594 p = 0.013*
VAS-H	A	6.74	1.86	7.00	1.54	− 1.347	0.191	− 0.3	n.s.
	B	6.988	2.0276	6.90	1.93	0.093	0.926	0.0
VAS-M	A	5.924	1.9528	7.04	1.83	− 2.562	0.017*	− 0.5	T * G
	B	5.780	1.4344	5.38	1.86	1.553	0.134	0.3	F = 8.710 p = 0.005**

Performance scores (means and standard deviations) at the pre-test and the post-test: higher scores indicate better performances with the exception of Stroop colour–word test (SCWT). Cohen’s d interpretation: small (0.2), medium (0.5), and large (0.8)

MMSE Mini Mental State Examination, CPM coloured progressive matrices, DSST digit symbol substitution test, DST digit span test, DSBT digit span backwards test, CT cancellation test, SCWT stroop colour–word test, PALT paired associate learning test, CBTT corsi block-tapping test, ROCF.a copy of Rey–Osterrieth complex figure B, ROCF.b immediate reproduction from memory of Rey–Osterrieth complex figure B, ROCF.c delayed reproduction from memory of Rey–Osterrieth complex figure B, COWA controlled oral word association, CFT category fluency test, HD humpty dumpty test, VAS-H health visual analogue scale, VAS-M mood visual analogue scale. T = time factor; G = group factor; T * G = interaction group × time

* < 0.05; ** < 0.01; *** < 0.001; n.s. = not significant

Performances in most tests (MMSE, CPM, DSST, CT, CBTT, ROCF.c, COWA, CFT, HD, and VAS-M) were significantly improved in the experimental group. In the control group, a significant difference was found only in the DST.

The Cohen’s d was moderate in MPC, DSST, CT, ROCF.c, COWA, HD, and VAS-M and large on CBTT in the experimental group. In the control group, there was only a small effect on DST.

In the repeated measures ANOVA, the interaction time and group significantly influenced the CT, CBTT, and VAS-M, while a marginal result was also found in ROCF.c. The group factor significantly influenced ROCF.a, and while the experimental group remained at the same level as measured in the pre-test, the control group exhibited lower scores. The time factor (T) significantly influenced DSST, COWA, CFT, and HD.

In terms of balance, despite a similar global cognitive performance (pre-test MMSE scores), the experimental group exhibited higher performances in attention, processing speed, spatial memory, and executive functions tests as compared to the control group. The control group exhibited a negligible improvement in the DST.

Furthermore, regarding the number of subjects who demonstrated positive differences between pre-test and post-test performances, the sign test only showed significant differences in the experimental group regarding the MMSE (p = 0.04), CBTT (p = 0.002), COWA (p = 0.007), HD (p = 0.02) and VAS-M (p = 0.007) while no significant differences were found in the control group.

Discussion

The growing interest of researchers aimed at identifying strategies useful in preventing or reducing the consequences of cerebral ageing is motivated by both scientific and socioeconomic reasons (Rosenberg et al. 2018). The conflicting data in the existing literature justify further research to establish if older people can benefit from these demanding strategies (Lustig et al. 2009).

The present research was inspired by a multimodal approach. It was planned to consider the assumed roles of physical, mental, and social factors: regular and self-regulated aerobic exercise simultaneously combined with a mental stimulus, hypothetically providing the ability to motivate learning about topics of personal interest, in addition to a series of conditions aimed at increasing social interaction.

Overall, the results obtained agreed with the abundant research that has reported the effectiveness of an active lifestyle in fighting the effects of cerebral ageing, delaying cognitive decline, and reducing the risk of dementia (Christie et al. 2017). Similarly, in regard to changes in mood, it is generally agreed that elders respond positively to a training programme (Diamond et al. 2015). A limitation to the generalization of the results is represented by the sampling methods, and volunteer subjects may have had peculiar characteristics such as sufficient motivation in committing to achieve a result or, more generally, in maintaining an active lifestyle. Moreover, it remains to be seen whether the positive effects are maintained over time: some studies have documented long-term effects, but in the majority of experiments, the evaluations were conducted immediately after training, a method which does not allow one to understand if the positive effects can be maintained after the training has stopped (Rebok et al. 2014). Moreover, the existing literature has not been able to document with certainty if the benefits gained could advantageously reflect on autonomy and independence in carrying out daily activities such as driving, shopping, paying bills and scheduling the taking of medications (Foubert-Samier et al. 2014). It is worth noting that the multimodal approach is thought to be effective in improving not only trained but untrained performances. The underlying assumption is that compared to formal training, multimodal interventions are not based on specific cognitive functional training, but on a change in lifestyle (Bielak et al. 2007; Christie et al. 2017). In summary, current studies provide consistent evidence for a protective benefit against cognitive decline offered by multimodal lifestyle interventions and suggest a positive influence on autonomy in everyday life (Hertzog et al. 2008). Nevertheless, further studies that directly examine and verify these changes through the appropriate scales of functional independence in activities of daily living and quality of life are necessary.

In our study, concomitant changes in physical activity habits, mental stimulation, and social participation positively influenced cognitive functioning and mood in healthy subjects aged ≥ 65. Above all, our findings suggest that even at an advanced age, cognitive reserves can be enhanced by relying on several different variables of the environmental context.

Two aspects must be emphasized when evaluating our results. First, the subjects of the two groups were matched not only by sex, age, and schooling but also by the perceived states of health and mood, as well as the amount of social interaction. In addition, the level of pre-test cognitive performance was the same in the two groups with the sole exception of ROCF.a, in which the experimental group showed a significantly higher performance. Therefore, the retest effect alone cannot justify this improvement in performance. Furthermore, the statistical analysis of individual performances carried out by the sign test suggests that the cognitive and mood benefits were actually achieved through the training.

Second, all subjects accomplished many tasks at higher levels with respect to their age. For example, in the backwards version of the digit span task (DSBT), while the mean value related to the age of our study groups is 3.92 (± 0.98), the subjects of our experimental group obtained a score of 4.42 (± 0.99) and the subjects of our control group 4.21 (± 1.02). It is possible that in some performances, the differences could be non-significant due to the ceiling effect, related to the normal values of the age group. Because typical ageing is characterized by progressive changes in cognitive function, and the effects of any treatment cannot be fully disclosed before physical or cognitive decline has become significant, it is possible that a study period of 6 months might not be sufficient enough to show significant cognitive changes. In other words, in assuming that the control group presents a reduction in performance while the experimental group maintains it longer or at a higher level, a longer observation time might amplify the differences between the two groups, even in those tests of our neuropsychological evaluation, in which currently the comparison is not significant. A longer follow-up is probably required to answer this question and verify this hypothesis.

Moreover, it is interesting to point out that this method includes mental, physical, and social activities completed at the same time and that all the participants did the training consistently, with the minimum amount of attendance being 80% of the 46 total hours of activity. Therefore, the proposed method seems to be considered enjoyable and to present the typical characteristics of leisure activities.

Overall, our findings can contribute further information about the effectiveness of a multimodal approach on cognitive functioning and mood in older subjects without cognitive impairment.