What is meant by chronological age?

1 The Adult Years

Some researchers suggest the use of chronological age as a marker for the timing of adulthood, whereas others suggest the transition to adulthood is better characterized by events or rites of passage such as graduation from school, starting a job, or having a family (Neugarten and Hagestead 1976). Adulthood is usually divided into several periods: young or early adulthood (approximately aged 20–39), middle adulthood (40–59), and old age (60+). Old age is typically divided into the periods of young old (60–75) and old old (75 and up).

Subjective definitions of age are also important. When adults are asked how old they feel, their responses often do not correspond to their actual chronological age. Those in adolescence often feel slightly older then their age, young adults usually report feeling close to their age, whereas in midlife and old age adults feel on average 10–15 years younger than their age (Montepare and Lachman 1989). The older one is, generally, the larger the discrepancy between age and subjective age. Feeling younger than one's age is typically associated with better health and well-being.

The social clock is another important organizing framework for adulthood. Based on cultural and societal norms, there is a sense of when certain events or milestones should be achieved. Thus individuals can gauge whether or not they are on-time or off-time relative to these norms (Neugarten and Hagestead 1976). There are consequences associated with being early or late with regard to certain events (graduation, marriage, having a child, getting a job). However, the research indicates that many individuals set their own timetables which do not correspond to societal norms. For example, those with more education will be likely to get married, start a family and begin a job at later ages than the general population. It is one's own constructed timetable that appears to be most important for well-being, and the social clock may be less critical.

2.3 Age Chronology

The clinical-developmental approach is far less focused on chronological age than are most research approaches to development and psychopathology. Unfortunately, it has become quite common to label a model or study as being an example of research in ‘developmental psychopathology’ if age is introduced as a significant variable. But chronological age is a very crude indicator of the underlying processes of biological maturation, cognitive capacity, peer and family relationships, and one's understanding of one's self. Age does, of course, play a role in development in most of these domains, but there is frequently a significant variation in maturity in each of these areas within a single age group.

There are, of course, also age trends in some forms of psychopathology. For example, suicide rates rise dramatically in adolescence, as do other disorders such as depression, and conduct problems. Schizophrenia is typically an illness that begins in late adolescence or young adulthood. Age, therefore, can serve as a simple way to organize an understanding of many underlying pathogenic processes. However, the simplicity of this approach is deceptive (see Rutter 1989).

In psychopathological development, as in normal development, we often find that chronological age is by no means a guarantor that basic cognitive and social-cognitive processes have occurred. For example, despite textbook claims to the contrary, many adolescents and adults never achieve formal-operational thought (i.e., tables in a typical psychology textbook list adolescence next to formal operations, as if all adolescents function at that cognitive stage). While many adolescents have achieved a formal operational level of thinking, others continue to function at the level of concrete operations, and indeed may remain at this level for the rest of their lives.

Consequently, if we use chronological age as the principle developmental marker, then the important variations that occur in every group of normally-developing individuals will receive insufficient attention. For the clinical-developmental psychologist, this suggests that chronological age needs to be supplemented with a domain-specific understanding of a person's capacities in different areas of development.

But, although capacities are often measured in terms of chronological age, the more sophisticated measurements of development that we have as we enter the twenty-first century call on us to subsume the idea of chronological age into our understanding of progressive adaptation and skill acquisition. In so doing, our research finds significant links between development and psychopathology in adolescents, where chronological age plays only a minor role. Level of social cognitive functioning and the shape of the internal self and object representations, rather than age, are often associated with specific disorders.

2 Markers of Old Age

There is no single criterion for defining old age. Though chronological age is not a good criterion to identify a person as ‘old’ or ‘aged,’ it is the one most commonly used. In Western countries, 65 years is considered as the cutting point between middle age and old age, in developing countries it is between 55 to 60 years. The basis for this is the average life span of an individual, which is linked with the retirement age. Apart from chronological age, there are other indicators/markers of old age. Biologically, menopause is considered the beginning of old age in the case of women. In Indian society, like many other traditional societies, marriage of one's own children and becoming a grandparent herald the beginning of gradual withdrawal from active occupational and family life and turning towards spiritual growth.

1 Introduction

Skeletal growth progresses with change in shape and size of the bones of the skeleton with age. Discordance in the bone age and chronological age further directs diagnostic evaluation of growth-related, endocrine, or metabolic disorders. Although bone age assessment remains the routine radiological procedure performed in pediatrics, the paradigm has not changed significantly since the work from Greulich and Pyle [1] or Tanner-Whitehouse [2] in which an atlas of radiographs of representative ages is used to compare the child's radiograph to determine the bone age, whereas the Tanner-Whitehouse method uses a scoring system by examining 20 specific bones to calculate bone age. Both techniques require a considerable amount of time and a significant amount of radiologist expertise, which further leads to intra- and interobserver variability resulting in contradictory line of action when bone age and chronological age discrepancies are encountered. This kind of diagnosis is highly subjective and qualitative due to sole human observation basis

Therefore, there's a compelling need to develop a computer-aided model for bone age prediction in order to quickly bring the cases of bone age discrepancies for the correct line of treatment. Deep learning approaches have attracted a lot of attention in the medical domain lately. However, these techniques rely heavily on a large amount of data to perform effectively, which is usually not the case in the medical image analysis domain. Our framework uses ideas from metric learning and learns new concepts from little data.

In this work, the sections are organized as follows. The importance and motivation behind this work is detailed in Section 2. Section 3 introduces the related fundamental work carried out in the past to deal with the problem of bone age assessment. The proposed methodology based on metric learning paradigm along with dataset and architecture is presented in Section 4. The hyperparameters used during training of the proposed model and its resultant visualization embeddings are presented in Section 5. Sections 6 and 7 cover the discussion and conclusion of the chapter.

Whereas it is obvious that an individual must have lived 100 years in order to be called a centenarian, there is little consensus about the chronological age associated with the average onset of old age. Demographic changes such as increased average life expectancy and the expansion of the population over age 60 in many countries lie behind suggestions that old age involves several life phases (e.g., a Third and Fourth Age). Functional decline is the expected trajectory of psychological change in old age. To understand the processes underlying change, it is important to distinguish between trajectories that are pathologically determined, death-related (e.g., terminal decline), and normative (i.e., usual and age-related). Researchers interested in successful aging and what might be possible in the future, given the evolution of a culture supportive of old age, examine individual difference and subgroup factors associated with relative maintenance of functioning, delayed decline, and longevity. Well-designed studies of representative samples of individuals in their eighties, nineties and centenarians are needed to provide information about the life history and psychological factors that contribute to a long life and to selective survival.

Introduction

Robots are increasingly becoming a part of our everyday lives—from the Roomba vacuum cleaner robot to manufacturing robots to cars that are increasingly automated. One exciting application of robotics is in health care and the home. This application is of increasing importance, as the proportion of older adults in our population increases. There are many older adults in need of help with daily tasks than human caregivers to provide such assistance. Robotics can fill this health-care gap and help older adults to age in place by increasing wellness, providing rehabilitation, and assisting with activities of daily life by compensating for functional age-related declines. There are many tasks that robots could be developed to assist older adults with. Today, there are three main categories of robots that can assist older adults to age in place. There are social companion robots, such as Paro, and tele-operated robots, such as telepresence. Finally, there are (semi-) autonomous assistive robots, such as the Roomba. These robots can assist older adults with a range of daily activities, such as fetching items, helping around the house, providing support in disease management, and facilitating social connectivity. However, for robots to be truly effective in helping older adults age in place, it is important to consider how these machines will fit into older adults' daily lives. To this end, the purpose of this chapter is to (1) provide an overview of the role robotics may play in aging in place; (2) categorize the types of robots that can support aging in place; (3) identify ways in which robots could assist with activities of daily living (ADLs), instrumental activities of daily living (IADLs), and enhanced activities of daily living (EADLs); and finally (4) discuss adoption and ethical considerations in applying robots to support aging in place.

7.1.1 Relationship of Population and Postcapture Ages

Carey and his colleagues first defined two age categories central to the captive cohort methods. The first is population age x which is referred to as the average chronological age of all individuals in the population at the time of sampling. The second is postcapture age y defined as the time interval between an individual's capture and its death. The concept underlying the captive cohort concept is based on the unique relationship between the age structure of a population and distribution of deaths of its members as is depicted in Fig. 8. Three aspects of the relationship between population age distribution and remaining times to death shown in this figure merit comment: (1) life span of the average individual in the young medfly population (top set of graphs in Fig. 8) will be relatively long and, therefore, distribution of postcapture ages of death will be skewed to the right. This is because only recently eclosed members have the potential to live to the most advanced ages. In contrast, postcapture life span of the average individual in an old medfly population (bottom set of graphs in Fig. 8) will be much shorter with no individuals living to the older postcapture ages; (2) the extremes in postcapture longevity distribution of sampled medflies shed important additional light on the presence of older and younger individuals in the population. The observations that some individuals are capable of living to the most advanced ages postcapture will indicate that new individuals are being added to the population, i.e., only newly eclosed medflies are capable of living to the extreme postcapture ages. Similarly, the observation that a fraction of sampled flies die relatively soon after they are captured reveals the presence of older and/or frailer flies; and (3) change in mean age of a population is inversely related to change in mean postcapture age. For example, an increase in population age implies that the average individual is older. It follows that these individuals have a shorter remaining days to live.

2.2 Cohort Differences in Cognitive Ability and Education

In addition to varying individual developmental trajectories, mental abilities also show different cohort trends as well. Some abilities show positive cohort trends with successive cohorts functioning at higher levels when at the same chronological age. Other abilities exhibit curvilinear or negative cohort trends. The two abilities showing the strongest positive cohort trends are inductive reasoning and verbal memory—both representative of fluid ability. Current cohorts of the elderly are thus at double disadvantage on these abilities due to relative early age-related decline on fluid ability, combined with strong positive cohort trends on these same abilities. More modest positive cohort trends have been shown for spatial and verbal abilities. In contrast, curvilinear cohort trends have been shown for numerical abilities with birth cohorts 1918–1920s showing higher functioning compared to earlier or later cohorts when at the same chronological age. There does however appear to be a slowing of these cohort differences, and it is estimated that during the first part of the twenty-first century the differences between cohorts will become smaller (Schaie 1996).

These cohort trends in abilities are multiply determined; however, increasing levels of education across cohorts as well as medical and health advances appear to have been strong influences. The impact of increases in educational attainment as well as shifts in educational practice toward discovery learning, procedural knowledge and metacognition may have contributed in particular to the strong positive cohort trends for inductive reasoning and verbal memory. A recent reduction in the magnitude of cohort differences in abilities may be related to a plateauing of the dramatic increases in educational attainment that occurred in the later part of the twentieth century. Alternatively, the slowing of cohort trends may reflect the decline in college-entrance exam performance reported for recent cohorts of young adults; these cohorts are now in their late twenties and thirties and are represented in longitudinal studies of adult cognition.

Types of Research Design

Figure 1 shows the way that the cohort study relates to other kinds of research design. Each of the large blocks represents a population or (subpopulation) of a country. The vertical axis of each block shows chronological age and the horizontal axis shows the time (or “period”) of data collection in chronological years. Each vertical section represents a population survey at a particular time point. Each horizontal section represents repeated surveys of a single age group across time—3 year olds, 10 year olds, and so on. Such “time series” of observations on one or more age groups or the whole population provide a means of monitoring changes in the population, indicating response to policy shifts or where policy development is needed. The Connecticut crackdown on drunk driving, which was accompanied by a fall in deaths from driving, was used by Donald Campbell as an example of a quasi-experiment in his classic paper, Reforms as Experiments.

Cohort Designs

The cohort study is represented by the diagonal line, which shows the age the cohort has reached at each year in historical time. Thus, a baby born in the year 2002 will be age 8 in the year 2010 and age 43 in the year 2045. The cohort is generally identified here with a particular age group of a population usually described as a birth cohort because all the members must share a common age. These can encompass all births over the given year, as in the British Millennium Cohort Study (see Fig. 2), or all births in a single week, as in the earlier 1946, 1958 and 1970 cohort studies. The U.S. National Children's Study focusing on the effects of environmental factors (biological, chemical, physical, and social) on children's health, and scheduled to begin in 2005, will be based on 100,000 births with follow-up over the next 20 years. Many cohort studies follow a wider range of groups including the birth year conceived either as a single cohort or as a set of birth cohorts. The variable intervals between follow-ups are also shown in Fig. 2, indicating the scope for cross-cohort comparison. Cohort studies in Scandinavian countries, such as Sweden, start at much later ages than birth, e.g., 10 or 13 years of age, relying on the comprehensive administrative data held on state registers to supply information about earlier ages.

Figure 2. British Birth Cohort Studies Programme. Note: 1. Initial survey carried out at ca. 8 weeks. 2. Initial survey carried out at ca. 9 months.

The cohort study can extend to multiple populations representing one or more countries, such as the Paths of the Generation Longitudinal Study established by Mik Titma following up high school graduates. In its second version, this extended from Estonia, where it began, to 15 countries and regions of the old Soviet Union—a sample size approaching 50,000 individuals. The study can also be based on the populations of different towns or regions within countries, such as the Dunedin Multi-Disciplinary Study and the Christchurch Child Development Study in New Zealand or the Malmo Longitudinal Study and the Gothenburg Evaluation through Follow-up Studies in Sweden, starting in 1938 and 1961 respectively. Alternatively, the study can involve follow-up of different groups (subpopulations), including those that have been subjected to some form of treatment or policy intervention, while others have not. Under “true” experimental conditions, there would be an allocation of individuals to such groups on a random basis, who then become the cohorts for investigation. In the absence of such allocation, i.e., the cohorts are naturally occurring, one has another form of quasi-experiment. The validity of such a design for experimental purposes depends on the extent to which the groups are matched with respect to all attributes related to the outcome.

The temporal sequencing of the longitudinal data that cohort studies produce offers a powerful means of control in comparison with a cross-sectional survey. One can monitor the circumstances and experiences through which cohort members' life histories unfold to assess how these interact with personal characteristics to produce particular social, economic, psychological, or health outcomes. Contemporary data accumulated over the life cycle—establishing the sequencing of events, circumstances, and characteristics—provide a particularly valuable tool in the search for causal explanations of differential development and its outcomes. Dated event histories can be used to describe and model dynamic processes, rather than the static states observed in cross-sectional “snapshots.” Such possibilities have been greatly enhanced by advances in statistical methodology for the analysis of longitudinal data and the development of high-powered computer technology to apply them.

Epigenetic clock

DNA methylation can be utilized to highly accurately predict age, also known as the epigenetic clock or DNA methylation estimated age (DNAmAge) (Horvath and Raj, 2018). DNAmAge is one of the commonly used techniques to estimate an individuals’ biological age, which in contrast to chronological age also reflects the aging processes within the individual. Several algorithms have been developed that calculate DNAmAge based on DNA methylation levels at pre-defined CpG-sites. For example, Hannum’s clock estimates DNAmAge based on 71 CpG-sites (Hannum et al., 2013), while DNAmAge using Horvath’s clock is assessed using 353 CpG-sites (Horvath, 2013). Of interest, both methods show a strong correlation > 0.9 between DNAmAge and chronological age. Additional age markers such as age acceleration can be calculated when subtracting chronological age from the estimated DNAmAge or biological age. Recently, more extensive models have been developed resulting in new epigenetic biomarkers for aging. For example, Levine et al. incorporated clinical measures of phenotypic age in their model resulting in the biomarker DNAm phenoAge (Levine et al., 2018). In addition, the biomarker DNAm GrimAge was developed based on DNA methylation estimators of plasma proteins and smoking pack years (Lu et al., 2019). Both new biomarkers show a strong association with lifespan and healthspan. While all these markers of age acceleration mentioned above have been associated with diseases such as Alzheimer’s disease and Parkinson’s disease as well as with mortality, only a few studies have investigated age acceleration and lung function (Marioni et al., 2015; Rezwan et al., 2020; Wang et al., 2020). Moreover, these studies only showed weak associations between the different measures of age acceleration and different measures of lung function. With the current concept of COPD being a disease of accelerated lung aging, it would be definitively worthwhile to investigate the association between age acceleration and COPD in further detail.