Depression and anxiety symptom networks across the lifespan

Abstract

Background: The relationship between anxiety and depressive symptoms is complex and may vary across the lifespan. Symptom network analyses offer a powerful tool to examine these interactions, but few studies have directly compared symptom networks in younger and older adults.

Methods: We analysed data from the Cambridge Centre for Ageing and Neuroscience (Cam-CAN) study, including 786 participants aged 18 to 88, who reported at least subclinical levels of symptoms on the Hospital Anxiety and Depression Scale (HADS). Network analysis was employed to examine symptom communities (clusters of related symptoms), within- and between-community connectivity (association strength), and centrality (symptom importance) across age groups.

Results: The overall network structure, separating anxiety and depressive symptoms into two communities, remained stable. However, older adults showed reduced connectivity within depression and between depression and anxiety. While 'panic' was a consistently central symptom, 'rumination' and 'restlessness' were the key bridge symptoms (i.e. linking anxiety and depression) in young and older adults, respectively.

Discussion: Our findings reveal both stable and dynamic aspects of depression and anxiety symptoms across the lifespan. Reduced within-community connectivity for depressive symptoms suggests greater heterogeneity in how depression manifests in older populations. The shift in bridging symptoms, from cognitive (rumination) in young adults to somatic (restlessness) in older adults, suggests subtle yet clinically important differences in how depression and anxiety are linked across the lifespan. Our findings support age-informed assessment and diagnosis of depressive and anxiety symptoms.

Keywords: ageing; anxiety; centrality; depression; network analysis; older people.

Figure 1

Depression-anxiety symptom networks in younger and older adults. (A) Network constructed using MI between HADS questionnaire items in younger adults, with edges representing the magnitude of MI associations between items. Edge thickness and colour (as indicated in the colour bar) indicate normalised MI. Node colours represent communities detected by the Louvain algorithm. (Blue shade: anxiety community, orange shade: depression community). (B) Same as (A) but for older adults.

Figure 2

Age-related differences in within- and between-community connectivity. (A) Violin plots displaying the distribution of within-community connection strength for each symptom within the depression (orange shade) and anxiety (blue shade) symptom communities, for both young (lighter shades) and older adults (darker shade). (B) Same as (A) but for between-community connections.

Figure 3

Comparison of centrality measures between age groups. (A) Strength centrality, representing the sum of connection weights for each symptom, is displayed for both young (grey line) and older adults (black line). Nodes are coloured according to their classification in the HADS, with anxiety symptoms (HA) in blue and depressive symptoms (HD) in orange. (B) Bridge strength centrality, which quantifies the importance of a symptom in connecting the depression and anxiety communities, is shown for young and older adults. Each node is represented by a grey line connecting its bridge strength values in the young (grey line) and older adults (black line) groups. Node colours correspond to their classification in the HADS, as in (A).