Are social isolation and loneliness associated with cognitive decline in ageing?

1Attachment theory suggests that human beings are born with a psychobiological system that motivates them to seek proximity to significant others in times of need. According to Bowlby, the goal of this attachment behavioral system, is to maintain adequate protection and support, which is accompanied by a sense of safety and security.

2The health, life and genetic legacy of members of social species are threatened when they find themselves isolated (see for a review). During the COVID-19 lockdown, for example, we experienced social isolation, and many people suffered the consequences of loneliness. During the pandemic, the dangers of the virus were prioritized. However, social isolation and loneliness can result in both short- and long-term health effects that cannot be ignored.; see for a review), for instance, showed the association of social isolation and loneliness with a significantly increased risk of death from all causes. Likewise, being disconnected posed comparable danger to smoking 15 cigarettes a day, and was more predictive of early death than the effects of air pollution or physical inactivity. also showed that people who had strong social relationships had a 50% increased likelihood of survival than those with weaker ties. Similar effects have been observed using data from the Framingham Heart Study, whereby chances of becoming happy, depressed, or obese were mirrored by similar changes in the closest friend. When friends where considered, each person was asked to name one friend, and not all of these nominations were reciprocated. Results showed that the contagion phenomenon was especially strong if the friendship was mutual (both friends nominated each other as friend).

3The topic of social isolation and loneliness is considered so important that, prior to the pandemic, the UK appointed a Loneliness minister and published a national strategy for tackling loneliness in 2018 and the World Health Organization declared that loneliness is a major health concern worldwide.

4In a recent review, summarize the evidence showing that loneliness impairs the immune system, thus reducing resistance to disease and infections. For instance, found that freshmen students who reported feeling lonely had a reduced immune system response when they were given a flu vaccine compared to students who felt socially well engaged. Moreover, those students with only four to 12 close friends had significantly poorer responses than those with more friends. Thus, feeling lonely and having few friends results in a particularly poor immune defense. also found that social bonds stimulate the release of the body’s natural killer cells, one of the white blood cells of the innate immune system whose core function is to destroy harmful bacteria and viruses. Finally, Cole and collaborators have shown that loneliness is associated with higher pro-inflammatory gene expression, indicating an upregulation of inflammatory signaling that can be a precursor for higher systemic inflammation and worse health. To conclude, the more immersed in a community—with social connections—the happier, and healthier, people are. Friends tend to act as our social support and are condition sine qua non for health quality.

5A second recent systematic review by has also shown abnormal brain structure (gray and white matter) and/or activity in the prefrontal cortex, insula, amygdala, hippocampus, and posterior superior temporal cortex associated with loneliness. Loneliness was also related to biological markers associated with Alzheimer’s disease pathology in two cross-sectional studies using PET imaging that found a significant relationship between loneliness and higher amyloid burden and greater tau pathology in right entorhinal cortex and right fusiform gyrus, especially in APOEε4 carriers.

6Both social isolation and loneliness refer to human connection and may have a relationship with cognition, but they are not synonymous. As stated by, the term “social isolation” reflects an objective reality, meaning a factual deficit in a person’s social bonds and support. On the other hand, loneliness refers to a subjective feeling of discrepancy between one’s wishes of social contacts and actual interactions.

7Research on social isolation over the years has shown that both objective (social isolation) and subjective (loneliness) components must be examined when investigating their association with health and wellbeing. They are two separate constructs that have shown only modest correlations (r ∼ 0.25–0.28;; also see;; ) and may have independent negative effects on older adults’ mental health. Also, individuals may experience loneliness without also suffering social isolation, or vice versa. This difference has been identified as social asymmetry by McHugh Power and collaborators.

8Older adults’ health and wellbeing may be specially threatened by social isolation. For example, they experience the loss of close others through ill health and bereavement, dislocation from their relatives due to increased familial mobility and greater difficulties engaging in social activity after retirement. Also, there is evidence that attachment style becomes more avoidant with age, especially when compared with younger adults. Knowing that cognitive functioning constitutes a major outcome of older adults’ health and wellbeing, the aim of the current review was to explore further this link between social isolation and loneliness and cognitive functioning in healthy older adults.

9Previous reviews have, however, addressed this question considering only separately the link between cognition and aspects of social relationships and between cognition and loneliness.

10There is however no published work considering both constructs as separate possible causes of cognitive decline. It was therefore important to investigate within the same study the specific contributions of social isolation and loneliness to cognition in older adults. The aim of the present review was to examine for the first time, in recent longitudinal studies (last 5 years), the relationship between social isolation, loneliness, as different concepts, and cognitive changes in later life.

11To do this, we followed Preferred Reporting Items Systematic Reviews and Meta-Analysis for scoping reviews (PRISMA-ScR) guidelines, with the pursue of achieving an appropriate organization and integrity of the work.

12The question we wanted to investigate was whether there is a link between social isolation, loneliness and cognitive decline among older people. In other words, do isolated and/or lonely older adults present higher rates of cognitive decline than people with rich/active social life/connections? If such relationship does exist, could it be bidirectional?

13To answer this question, the following inclusion criteria were established: studies (1) that examine the relationship between social isolation, loneliness and cognitive function, (2) using a longitudinal design, (3) including participants with a mean age ≥ 60, (4) with no diagnosis of cognitive impairment nor dementia, (5) reporting original data, (6) published in English in a peer-reviewed journal, (7) released between January 20171 and April 2021. Reviews or opinion papers were excluded.

14Pubmed, Scopus, Web of Science (WOS), Medline and PsycInfo were the selected electronic databases through which the literature was searched. The search was conducted on April 2021, and it included the following terms: “social isolation,” “loneliness,” “ageing,” “older adult,” “elderly,” “cognition,” “cognitive function,” “cognitive processes,” “cognitive control,” “executive control,” “executive function,” in various combinations. For instance, in Scopus the following search was performed: (“social isolation”) AND (loneliness) AND (aging OR ageing OR “older adults” OR elderly) AND (cognition* OR “cognitive function*” OR “cognitive control” OR “cognitive processes*” OR “executive control” OR “executive function*”) AND (LIMIT-TO (PUBYEAR, 2021) OR LIMIT-TO (PUBYEAR, 2020) OR LIMIT-TO (PUBYEAR, 2019) OR LIMIT-TO (PUBYEAR, 2018) OR LIMIT-TO (PUBYEAR, 2017) AND (LIMIT-TO (LANGUAGE, “English”).

15As shown in Figure 1, the initial search yielded 2,673 results, from which 2,262 articles were identified in Scopus, 216 in PubMed, 116 in WOS, 63 in Medline and 31 in PsycINFO. A systematic filtering process was then conducted by the two authors, and in case of uncertainty, the final decision on inclusion was made through discussion.

16First, 206 duplicates were removed. Titles and abstracts—and when in doubt also the methods—from the remaining 2,467 unique records were then screened, leaving 23 articles that were fully read and examined to determine eligibility for inclusion. Upon review, the final sample of this study is composed of 12 scientific publications (see Table 1). The 11 studies that were excluded for not fulfilling any of the criteria are listed in Table 2 along with the reason for exclusion.

17The following information from each of the included studies was extracted (see Table 1): authors and year of publication, country and setting in which it was carried out, objective of the investigation, study design, characteristics of the sample, loneliness and/or social isolation measurement, cognitive domains measurement, covariates, and results obtained.

18Only longitudinal studies were included in this scoping review, as they represent the best way to investigate associations between risk factors and cognitive decline and try to establish directional relationships (e.g., ). Follow-up periods for all 12 longitudinal studies ranged from 3 to 11 years. All studies reported findings from participants enrolled in large, population-based studies, being these the Survey of Health, Ageing and Retirement in Europe (SHARE; ), the English Longitudinal Study of Ageing (ELSA;; ), the Cognitive Function and Ageing Study-Wales (CFAS-Wales; ), the Cambridge City over-75s Cohort Study (CC75C; ), the Irish Longitudinal Study of Ageing (TILDA;, ), the Collaborative Research on Ageing in Europe study in Europe (COURAGE; ), the Health and Retirement Study (HRS; ), the Chinese Longitudinal Healthy Longevity Survey (CLHLS; ), the China Health and Retirement Longitudinal Study (CHARLS; ) and the German Ageing Survey (DEAS; ).

19Eight studies were conducted in Europe, including Germany, Spain, Ireland, and the UK. Two studies were carried out in China and one in the USA. Lastly, sample was formed using data from several countries—Austria, Belgium, Denmark, France, Germany, Greece, Italy, Spain, Sweden, Switzerland, Netherlands, and Israel.

20In all studies samples included both genders, and the number of participants per study ranged from 713 to 14,114 for a total of 77,265 participants across all studies. All studies’ participants were independently living older adults. Most participants were described as generally healthy, although some suffered from chronic physical diseases such as hypertension, diabetes, cardiovascular conditions, arthritis as well as psychological disorders—such as depression and anxiety.

21From the nine studies investigating loneliness, some utilized the UCLA loneliness scale or a three-item shortened version of it (the Hughes loneliness scale,;;;; second wave;; ). This Likert style questionnaire was designed to measure general feelings of dissatisfaction with one’s social interactions, and it includes questions such as “How often do you experience loneliness?” The other researchers used single questions. and used yes/no format questions such as “do you feel lonely?.” single—item question was part of the abbreviated version of the Center for Epidemiological Studies Depression scale (CESD; ). Participants were asked: “How often have you experienced the following feelings over the last week: I felt lonely?.” Response options were: 1 = Almost all of the time; 2 = Most of the time; 3 = Some of the time; 4 = Almost none of the time. also used this single question from the CESD.

22Six studies investigated social isolation. From these, assessed perceived social isolation using the scale, which includes 4 items. Each item ranges from 1 (strongly agree) to 4 (strongly disagree). Some other authors used indexes. For example, and used an index based on the one proposed by, including 5 binary questions such as “did the responded live alone?,” with scores ranging from 0 to 5, and higher scores representing greater social isolation. combined three items to create an index of social isolation, adapted from previous research. One point was assigned, for example, if participants were not married or had less than weekly contact (by phone, in person, or by e-mail) with children. Scores of social isolation ranged from 0 to 3, with higher scores indicating greater isolation. Finally, used the Lubben Social Network scale-6 (LSNS-6, ), a self-report standardized measure of social engagement including family and friends, constructed of three sets of questions. The three items assess the number of relatives/friends the participant sees or hears from at least once a month, could call on for help, and can speak with about private matters. Responses are collected using a six category response, in which the participant indicates the number of relatives/friends available. Response scores range from 0 (no relatives/friends) to 5 (nine or more relatives/friends). The overall scores for each six questions are summed and range from 0 to 30, with higher scores indicating lower social isolation. A score of ≤ 12 may be taken to indicate the presence of social isolation. used a scale to measure frequency of three types of contact with social network. Participants had to rate how often they (a) met up (include both arranged and chance meetings), (b) spoke on the phone, (c) wrote to or emailed their children, other family members, and friends, respectively. Options included three or more times a week, once or twice a week, once or twice a month, every few months, once or twice a year, less than once a year, or never.

23Episodic memory was assessed in different studies via word recall. For example,;, and used the 10-word list immediate and delayed verbal recall from the CERAD. In terms of attention, used the digit symbol test adapted from the digit symbol substitution test. used the Sustained Attention to Response Task (SART, ), where a series of single digits between 1 and 9 are sequentially presented and participants are instructed to press a keyboard key as soon as possible (with response time noted) for each digit presented, except if the digit is 3. used the digit span from the Wechsler Adult Intelligence Scale. Verbal fluency was also assessed by;,, and. Other studies assessed global cognitive function using the Mini Mental State Examination (MMSE,;; ). It is noteworthy that this test has been criticized and is being replaced by the Montréal Cognitive Assessment (MoCA) as it is more sensitive and specific to detect early cognitive impairment than the MMSE. used the Cambridge Cognitive Examination (CAMCOG), a standardized instrument used to measure orientation, language, memory, praxis, attention, abstract thinking, perception, and calculation. Finally, used the Modified version of the telephone interview for cognitive status (TICS; ), which includes questions of orientation, repetition, naming, and calculations.

24Six studies from our sample examined the relationship between social isolation and cognitive functioning., after examining a German sample of 6,420 people, concluded that increases in social isolation were associated with decreases in cognitive function, specifically perceptual motor speed and processing speed. Likewise, revealed a link between social isolation and memory decline in 11,233 later-life adults in England, suggesting that the former seems to affect the latter, but not the reverse. The study by explored this relationship in 2,224 Welsh older adults, assessing multiple domains and showed that being socially isolated might lead to poor cognitive functioning. This research also analyzed the role that cognitive reserve could play in the relationship between social isolation and cognition in later life, showing that, although no link was found when explored cross-sectionally, at 2-year follow up, cognitive reserve seemed to moderate the association. Particularly, the association between social isolation and cognitive change was non-significant for those participants whose former main employment was considered complex (i.e., doctor or lawyer) and those whose social and economic class were higher.

25examined data about objective social isolation, loneliness and recall from 6654 US individuals during 6 years with follow-ups every 2 years. Concerning social isolation, their results suggest that it is a predictor of lower cognitive functioning and it accelerates cognitive decline. In the same way, after analyzing data from 7,761 Chinese adults over 50 years, concluded that social isolation was significantly linked to cognitive decline in all cognition measures (i.e., episodic memory, orientation and visuospatial and numeric ability) at follow-up, even after taking into account loneliness and other covariates such as chronic diseases or depression. evaluated 1,691 Spanish participants aged 50 or older to determine whether loneliness and/or social isolation were related to changes in cognition over a 3-year follow up period. They found that individuals with a higher isolation index tended to present lower scores in verbal fluency, forward digit span and in the composite cognitive score.

26Nine studies from our sample investigated the association between loneliness and cognitive function. Some reported that loneliness might be a predictor of cognitive changes over time, while others did not. The study by suggests a bidirectional relationship between both domains and found that attention may affect loneliness and not reversely. Finally, concluded that the link between loneliness and cognitive decline over time was significant among men but not women.

27study concluded that loneliness had a significant association with lower scores in immediate and delayed recall, verbal fluency, backward digit span and the composite cognitive score that evaluated overall cognition. The research by is the study with the largest sample (14,114 individuals assessed over 11 years) and culturally the most heterogeneous (comprising 12 European countries). It evaluated memory recall and verbal fluency, and the results suggest that feeling lonely increases the risk of suffering cognitive impairment regardless of all covariates controlled (i.e., social isolation, depressive symptoms).

28Moreover, found that loneliness had a significant association with a decline in episodic memory, orientation and visuospatial and numeric ability before controlling for variables such as chronic diseases, health behaviors, disabilities, and depression. However, after controlling for them, this relationship became insignificant. In the same vein, reported that, unlike social isolation, loneliness does not seem to predict changes in cognition longitudinally, although it correlated with lower cognitive function cross-sectionally. study is the only one whose sample is composed solely by individuals aged 75 or older evaluated over a 20-year period, reason why its sample (713) is the smallest. The results showed that loneliness did not seem to be a significant long-term harmful risk factor for cognitive decline among the older old.

29Finally, focused their efforts on analyzing whether the association between loneliness and cognitive function varied among 6,898 Chinese men and women 65 years and over. Notwithstanding older women reported feelings of loneliness more frequently, the repercussion of loneliness on cognitive decline over time was significant among older men but not women.

30A remaining question from previous reviews was whether the association between social isolation and cognition or between loneliness and cognition might be bidirectional. The fact that the reviewed studies in our scoping review were longitudinal may contribute to answer this question.

31When analyzing the association between loneliness and cognitive function (memory and verbal fluency), observed that higher loneliness predicted poorer memory and verbal fluency at baseline and influenced such cognitive domains in a negative way after a 10-year follow up. Besides, worse baseline memory (but not verbal fluency) and its rate of decline over time seemed to contribute to an intensification of loneliness at follow-up. Additionally, higher baseline memory seemed to predict a slower change in loneliness, revealing a bidirectional relationship among loneliness and memory. In another study, found that cognitive functioning at wave 1 predicted loneliness at wave 3 and loneliness at wave 1 predicted cognitive functioning at wave 3. Finally, also hypothesized a bidirectional relationship between loneliness and sustained attention. Nevertheless, sustained attention at baseline predicted loneliness at 4-year follow-up, but reverse results, loneliness predicting sustained attention, were not found. This study is unique in that it is the only study in which solely cognitive function prognosticates loneliness longitudinally.

32Five out of the 12 reviewed studies considered the role of depression as a possible mediator between social isolation, loneliness and cognitive decline. In the study by depression in the previous 12 months was assessed with an adapted version of the CIDI 3.0. concluded that the results were not affected by depression, as they remained unchanged after excluding individuals with depression. Similarly, reported that, albeit loneliness and depression measured using seven items of the CES-D, seemed to be intimately linked, ultimately, they were independent and loneliness might be associated with memory decline over a decade by itself, despite of depressive symptoms.

33Three studies showed however a significant link between loneliness and depression. The first showed that higher levels of depressive symptomatology (measured using the validated and reliable 20-item CES-D scale, ) predicted higher levels of loneliness and worse performance on sustained attention. also observed that depressive (20-item CES-D scale), but not anxiety, symptoms mediated the relationship between loneliness and cognitive functioning. Finally, demonstrated that the association between loneliness (not social isolation) and cognitive decline became insignificant after depressive symptoms (measured with the 10-item CESD-10, ) were taken into account.

34The aim of this scoping review was to investigate for the first time possible associations between social isolation, loneliness, as separate constructs, and cognitive function in cognitively healthy older adults. To do this, we reviewed longitudinal studies investigating social isolation, loneliness and multiple cognitive functions. Studies were carried out in three continents (Europe, Asia, and America) and included large samples with a total of 77,265 participants across studies.

35Results from the selected studies revealed that social isolation, loneliness and cognition are related. The relationship between social isolation, understood as an objective manifestation of lack of social bonds and support, and cognition seems robust, with results congruently (6 out of 6 studies) showing that it is negatively associated with cognitive functions in older populations. Previous reviews assessing studies published before 2017 had shown some heterogeneity in the results, depending on variations in approaches to measuring social activity and social networks across studies. Our scoping review on recent longitudinal studies however adds consistent evidence to the idea of a protective effect of social connection through life on cognitive functions.

36The results of the recent studies looking at the effect of loneliness on cognitive function seem to be less consistent. Despite the common knowledge that “loneliness kills,” the relationship between loneliness, understood as the subjective feeling of discrepancy between one’s wishes of social contacts and actual interactions and cognition is only clearly shown in 4 out of 9 studies. While some studies identify loneliness as a predictor of cognitive function, others do not or show it only for men. Moreover, this relationship may be, at least partially, mediated by depression.

37Measuring loneliness is not a simple issue. It was measured by the UCLA loneliness scale in one study and by a three-item shortened version of it in five studies (;;; second wave;; ). The other researchers used single questions. As an example of how prevalent this approach is, from the 10 studies examining loneliness in the ageing population in meta-analysis, seven of those made use of single-item questions to measure loneliness while the remaining three used either the 3-item, short form of the Revised-UCLA Loneliness Scale or the six-item De Jong-Gierveld Loneliness Scale. Also, recent research has shown that the subcomponents of loneliness scales were notably decoupled by the confinement during COVID, supporting the notion that loneliness is not a unitary, isolated construct but rather represents a cluster of different experiences of social integration and socioemotional states. It is then possible that to capture feelings of loneliness and detect associations with cognitive decline, a full scale should be used (as was the case only in the study by ).

38define loneliness as a signal that one’s connections to others are weakening and to motivate the repair and maintenance of connections to others that are needed for our health, wellbeing and survival of our genes. This definition considers that loneliness evolved to improve survivability when socially isolated, through hypervigilance and increasing motivation to connect with others (also see ). Putting together the results from the current review and previous studies, the findings of depression as a mediator between loneliness and cognitive decline found in some studies would be compatible with this conception. Importantly, loneliness has been previously found to be closely associated with depressive symptoms. The link between depression and loneliness is clearly established by authors that conceive depression as an evolutionarily conserved mechanism to terminate separation distress (;;; also see; social safety theory), placing social loss at the center of the neurobiological dynamics (e.g., inflammation, altered HPA axis functioning, declining neurotrophins prosocial peptides and amines) driving depression. Also, two studies excluded from this scoping review (see Table 2) demonstrated the key role played by depression as a mediator between loneliness and cognition. study with 8,382 USA participants aged 65 or older concluded that, after 12 years, the effect of loneliness on cognitive function became marginally significant after controlling for depression, and that the latter is associated with a more rapid cognitive decline. Similarly, revealed that, only among the individuals who reported higher levels of depressive symptoms, loneliness was associated with poorer cognitive function.

39Interestingly, this link with depressive symptoms was not observed for social isolation. In the scoping review by on social isolation, loneliness and health in old age, 25 studies looked at the link between loneliness and depression and 3 looked at the link between social isolation and depression. The evidence reviewed clearly showed that loneliness is a strong risk factor for depression in old age, even after controlling for a number of covariates such as demographic characteristics, marital status, social isolation and psychosocial risk factors. The detrimental effect of living alone on depression was more often due to loneliness for men than for women. The evidence for a link between social isolation and depression was however weaker. This evidence suggests that the subjective experience of loneliness may be more strongly related to depression than the objective isolation per se, and that the link between social isolation and cognitive dysfunction is less mediated by depression.

40Considering the studies that investigated social isolation and loneliness simultaneously, and results also suggest that social isolation may have a more substantial impact than loneliness on cognitive decline. For instance, correlations between loneliness and cognitive function were non-significant after controlling for a wide range of demographic and psychosocial risk factors thought to influence loneliness. Consistent with our findings, recent research has suggested that social isolation is more associated with objective cognitive impairment outcomes, while loneliness is more associated with a subjective dimension of cognitive function (; see Table 2). This is also consistent with review showing that, when social and emotional loneliness were examined, the first seemed to have a stronger correlation with global cognitive function.

41These findings suggest that both social isolation and loneliness may impact cognitive health but probably in a different way, with a stronger weight on cognitive reserve for the first and on emotional wellbeing for the second. A possible explanation of the link between social isolation and cognitive decline would have to do with the “use it or lose it” perspective. reviewed the evidence suggesting that engagement in an environment that requires cognitive effort may facilitate cognitive function in older adults. This view, linked to the notion of neuroplasticity (also see ) suggests that the brain can be conceived as a muscle and that engagement in intellectual, social and physical activities stimulates the brain. If engagement in everyday activities is absent, it may result in disuse of the brain which will result in decline of cognitive functions. The seemingly stronger association between cognition and social isolation than between loneliness and cognition would go in favor of this explanation.

42On the other hand, although the empirical literature regarding the underlying biological mechanisms involved in social isolation and loneliness is scarce and not fully consistent, there is some evidence suggesting that, biologically, loneliness may trigger immune system impairment, chronic inflammation, hypertension, hyperlipidemia, hypercortisolism and prolonged activation of the hypothalamic–pituitary–adrenal (HPA) axis, which in turn could lead to a decrease in dendritic arborization in the prefrontal cortex (CPF) and hippocampus. Nonetheless, many of these neurological and physiological changes are linked to ageing itself, and, hence, the mixed effects of loneliness, social isolation and ageing may be even more complex.

43As for the direction of the relationship, the idea of this review came from the interest in the consequences of the isolation experienced during the COVID-19 pandemic, with the question of whether social isolation might be detrimental to cognition. In that sense, the direction of the effect may be conceived from social isolation and loneliness to cognition. However, there is evidence suggesting that the effect can also go in the opposite direction, i.e., poor cognition can lead to isolation and loneliness, known as reverse causality. It has been suggested that older adults who experience declining self-efficacy and loss of attachment relationships defensively place more emphasis on independence and self-reliance and less on interdependence. reached such a conclusion following their result of a higher rate of memory decline predicting loneliness at follow-up and vice versa. Possible explanations for this is that poor cognition may generate feelings of insecurity., for example, showed that decreased verbal fluency was a significant predictor of social loneliness. Poor communication skills may discourage conversation, hinder meaningful relationships, and thus increase loneliness. Another recent study (; see Table 2), also found that a dysfunction in working memory and planning might prognosticate higher perceived loneliness. also revealed this reverse association, demonstrating that sustained attention at baseline predicted loneliness 4 years later.

44Previous reviews had mentioned the difficulty to solve this question due to the great proportion of cross sectional studies (see for example,; ). Our review minimized this risk by excluding participants with cognitive impairment of dementia and by excluding cross-sectional studies. To conclude, it is likely that the relationship between social isolation loneliness and cognitive decline may be bidirectional.

45As limitations of the present study, the heterogeneity observed in measures of social isolation, loneliness and cognitive functions in the reviewed studies makes it difficult to carry out more specific analyses (e.g., meta-analysis). Also, only 12 studies fulfilled the inclusion criteria. As strengths, only powerful longitudinal studies evaluating cognitively healthy older participants were reviewed. This is the best method to explore causal relationships between the studied variables. Moreover, studies from three different continents were included, which helps generalizing the results. Finally, a key strength is that we considered both social isolation and loneliness as measures of objective and subjective isolation, respectively, which should contribute to a better understanding of differential effects of both constructs on cognitive decline in ageing.

46After considering the 12 longitudinal studies that comprise large heterogeneous, and culturally diverse populations, we conclude that both loneliness and social isolation, common among older adults, may be associated with cognitive decline. It seems that the relationship between social isolation and cognitive decline may be stronger than the link between loneliness and cognition, possibly mediated by depression.

47We also notice that both constructs are complex, with a diversity of definitions and measurements, and require more nuanced examination, with special attention to the specific nature or forms of social isolation, loneliness and their interactions in affecting mental health and cognition. Additional research is necessary to determine more precisely the causality and biological mechanisms implied in the association between social isolation, loneliness and cognitive functioning.

48The original contributions presented in this study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

49Both authors listed have made a substantial, direct, and intellectual contribution to the work, and approved it for publication.