Elsevier

Physiology & Behavior

Volume 70, Issues 1–2, 1–15 July 2000, Pages 49-53
Physiology & Behavior

Articles
Sleep estimation from wrist activity in patients with major depression

https://doi.org/10.1016/S0031-9384(00)00228-6Get rights and content

Abstract

Actigraphy has been used to monitor individuals' sleep and wakefulness patterns without laboratory confinement. To date, its validity in monitoring sleep and wakefulness among patients with major depressive episodes has not been systematically examined. The present study investigated whether the normative criteria of the Actigraph Data Analysis Software, initially optimized for healthy individuals, could score wrist-activity data accurately in a sample of depressed patients. Application of the normative algorithm yielded a correlation coefficient of 0.85 and an average error of 35 min, comparing actigraphic and polysomnographic sleep estimates. The algorithm optimized for this sample provided a correlation coefficient of 0.81 and an error of 6 minutes. For both algorithms, agreement for individual comparisons varied substantially. These findings suggest that scoring criteria optimized on wrist-activity data of healthy young adults may not produce optimal results for patients characterized with major depressive episodes.

Introduction

The need for less intrusive and cost-effective measurement methods has led to numerous new methods in clinical medicine as well as epidemiological inquiries. Establishing the validity of those methods remains a fundamental and challenging task to assure standardization and interpretation of data across laboratories or from differing methodologies.

The actigraph (ACT) methodology allows continuous monitoring of individuals' sleep and wakefulness patterns without laboratory confinement. Several investigations have established the validity and reliability of actigraphy in healthy persons 3, 8, 17, 19. It remains uncertain, however, whether actigraphy is valid and reliable for persons of various ages or with different clinical diagnoses 3, 5, 13.

In a validation study conducted with a sample of healthy individuals and patients exhibiting psychiatric illnesses, alcoholism, chronic pain syndromes, and other complaints, investigators found a lower validity coefficient for patients versus nonpatients (r = 0.81 and r = 0.92, respectively) [13]. In a similar study, Cole et al. [3] observed that algorithms developed for a mixed sample (i.e., healthy adults, healthy elderly, or psychiatric and sleep apnea patients) might not perform equally well for all subsamples. Even within a sample of patients diagnosed with insomnia, differing validity indices have been found for different diagnostic subclassifications 5, 9. These findings suggested that optimized criteria might be necessary to detect sleep and wakefulness accurately from actigraphic data of characteristically different samples.

Although actigraphy has been used in several studies to measure sleep patterns of depressed patients [20], its validity has not been systematically tested against polysomnography (PSG) in individuals with major depression. Accurate detection of sleep from actigraphic data of depressed persons might require optimized criteria to reflect reduction of activity level 10, 11, 20. Furthermore, because actigraphy relies solely on wrist motion, it might overestimate sleep time in depressed individuals who lie in bed motionless while awake. In those cases, the usefulness of actigraphy remains uncertain.

The present investigation assessed the validity of scoring criteria of the Actigraph Data Analysis Software (ADAS), initially optimized for healthy young adults, when applied to wrist-activity data of depressed patients. Moreover, we examined whether new criteria could be developed to optimally characterize sleep and wakefulness patterns of those patients. Derived actigraphic sleep estimates were referenced to polysomnography, the gold standard.

Section snippets

Participants

Twenty-four patients (mean age = 45, SD = 9, men = 17, women = 7) were recruited from ongoing polysomnographic sleep studies of individuals who met diagnostic criteria for a current Major Depressive Episode. Diagnoses were made on the basis of the Structured Clinical Interview for DSM-IV (SCID-IV); 18 patients received a diagnosis of insomnia, and 6 were hypersomniacs. Six volunteers were inpatients, and 18 were outpatients. Whereas volunteers (n = 19) had been unmedicated, 5 patients had been

Linear relationships between ACT and PSG sleep estimates

Results of the comparisons between actigraphic parameters (TST and SE) and corresponding PSG parameters are provided in Table 2, using the normative criteria. Relative to PSG estimates, average actigraphic estimates of TST (d = 35 min, confidence interval, 25–45 min) and of SE (d = 8%, confidence interval, 6–10%) were relatively high.

When actigraphy was used to estimate polysomnographic TST in a linear regression analysis, we observed a standard error of 19 min, with 72% of the variance in PSG

Discussion

Actigraphic studies of depressed patients have provided a wealth of data on circadian rest–activity rhythms, daytime activity level, and responses to different treatment manipulations 10, 14, 20, 21, 24. To date, however, research has not systematically addressed the need for optimization of scoring algorithms for depressed patients, although it is known that criteria developed for healthy persons do not yield optimal results when applied to clinical samples 2, 5, 9, 23. Furthermore,

Acknowledgements

This research was supported by the NIMH (MH30914-21, 57134-04, and 18399) and the NHLBI (HL55983 and HL40930). We thank Gavin Christie, Anna Demodena, Kathy Resowsky, and Don Sweat for their assistance in this study.

References (24)

  • G Jean–Louis et al.

    Actigraphic assessment of sleep in insomniaapplication of the Actigraph Data Analysis Software (ADAS)

    Physiol Behav

    (1999)
  • H Kuhs et al.

    Psychomotor activity in unipolar and bipolar depressive patients

    Psychopathology

    (1992)
  • Cited by (54)

    • Agreement between actigraphic and polysomnographic measures of sleep in adults with and without chronic conditions: A systematic review and meta-analysis

      2019, Sleep Medicine Reviews
      Citation Excerpt :

      The correlations for the TST measures ranged from .19 to .98; SOL correlations varied from −.07 to .80, while the range of correlations for SE was −.18 to .89, and the range for WASO was −.69 to .82. Fourteen studies included a total of 724 (17.5%) participants [M age = 50.9 (SD 19.5), 62.6% female] who had chronic conditions, including low back pain [18], dementia [10], bipolar disorder [19,34], schizophrenia [19], atopic dermatitis [20], depression [22,64], chronic fatigue [23], type 1 diabetes [25], heart failure [31], chronic obstructive pulmonary disease [33], Creutzfeldt-Jakob disease [39], Parkinson's disease [43,48], Huntington's disease gene carriers [45], depression and insomnia [47], and fibromyalgia [51]. Among these participants, ACT overestimated TST, compared to PSG [M = 22.42 min (CI 95%: 11.92, 32.91 min) n = 648 participants] and SE [M = 5.21% (CI 95%: 1.41%–9.00%) n = 622 participants], while ACT underestimated SOL [M = −7.70 min (CI 95%: −15.22, −0.18 min) n = 484 participants], and these were statistically significant.

    • Racial/ethnic disparities in sleep health and health care: Importance of the sociocultural context

      2015, Sleep Health
      Citation Excerpt :

      Other disparities in race-gender interactions on sleep efficiency and latency were also noted. A similar observation was made in a study conducted by one of the authors (GJL) in San Diego where blacks reported nearly 1 hour less of nocturnal sleep than whites.51 A meta-analysis of 14 studies conducted in the United States with > 4000 participants included a wide range of studies including those that used polysomnography to assess sleep architecture.

    • Sleep in America: Role of racial/ethnic differences

      2013, Sleep Medicine Reviews
      Citation Excerpt :

      Previous data from a study using telephone interviews document a greater prevalence of short sleep among blacks compared with whites living in Detroit, Michigan (18.7% vs. 7.4%).22 Actigraphic data obtained from adults residing in San Diego showed that white individuals slept an average of 6.32 h, whereas black individuals averaged 5.9 h a night.23 A similar observation was made in a sample of Chicago residents, for whom actigraphic sleep duration among white women was 6.7 h; white men, 6.1 h; black women, 5.9 h; and black men, 5.1 h.7

    • Actigraphy monitoring of symptoms in patients with Parkinson's disease

      2013, Physiology and Behavior
      Citation Excerpt :

      Our correlational analyses further show that changes in these movement symptoms were related to changes in UPDRS part II and III scores, which reflect severity in day-to-day activities and motor severity. Although there has been an increasing use of actigraphy to quantitatively evaluate symptoms of PD, such as tremor and sleep disorders [4-7] and other chronic disease conditions [13–16], few studies have undertaken evaluation of PD severity by actigraphy [7]. Thus, this study contributes to a deeper understanding of actigraphy and monitoring of symptoms of PD.

    View all citing articles on Scopus
    View full text