Mild cognitive impairment and subjective cognitive impairment


Masatoshi Takeda MD, PhD, Department of Psychiatry, Osaka University Graduate School of Medicine, 2-2 Yamadaoka, Suita, Osaka 565-0871, Japan. Email:

Until a decade ago, the common understanding was that forgetfulness in the elderly is a physiological phenomenon and that all elderly people show memory loss to some extent, with considerable variations in the severity of memory loss between individuals. Commonly observed memory loss in the elderly was termed ‘benign senescent forgetfulness’,1 which was regarded as different from the pathological state of dementia. Kral described benign senescent forgetfulness as follows: (i) impossible to recall unimportant things and part of the experience; (ii) able to recall these forgotten things at other times and on other occasions; (iii) more memory loss with older things compared with recent things; and (iv) subjects recognize this forgetfulness and try to compensate for it by several means.2 In the 1980s, there was an interest in ‘age-associated memory impairment’ (AAMI), which was also regarded as different from Alzheimer's disease, in which only memory function is disturbed without an affect on other cognitive functions.3,4 In those days, memory loss in the elderly was regarded as common and different from the pathological conditions leading to dementia, including Alzheimer's disease.

Recent findings in neuroscience do not necessarily support the notions described above, and it is conceived that memory loss in the elderly is not inevitable, because newer findings show fully functioning brain activity in the elderly. For example, there is no significant decrease in the number of neurons in the elderly brain,5 even though the size of the neurons is smaller, and there is no reduction in cerebral blood flow and cerebral glucose metabolism in the brain of healthy elderly people compared with the younger brain.6 Neurogenesis is observed in the dentate gyrus of the hippocampus of brains over 50 years of age.7 We have now abundant data that support the hypothesis that the human brain can function properly over 50 years of age if the elderly person is not affected by dementia.

Furthermore, this new understanding has been accelerated by research into the pathogenesis of Alzheimer's disease that explores possible means of intervention and prevention. Since 1996, drugs for the symptomatic treatment of Alzheimer's disease have been used widely, including acetylcholinesterase inhibitors (donepezil, rivastigmine, and galantamine) and an N-methyl-d-aspartate (NMDA) antagonist (memantine), and we are now at the stage of developing disease-modifying drugs, which will give more benefit to the patients if administrated in earlier stages of the disease. Considering these situations, the initial stage, or even prestage, of the disease attracts more attention in research because it would be the period when the disease-modifying drugs would exert their maximum effect.


The onset of neurodegenerative dementia, including Alzheimer's disease, is usually insidious; often, it is not possible to determine the exact time of onset. The diagnosis of dementia is often given to the patients retrospectively after identifying a certain set of behaviors caused by cognitive impairment. There is a relatively long period when patients are somewhat impaired but have not yet fully developed the symptoms of dementia. The prodromal stage of dementia has been described by many clinicians, and scales to assess the prodromal stage of dementia are used in clinical settings, such as the Clinical Dementia Rating Scale (CDR) developed by Hughes et al.8 and the Global Deterioration Scale (GDS) developed by Reisberg et al.9 The CDR breaks down the span between the healthy brain to severe dementia into five stages: 0, normal healthy; 1, mild dementia; 2, moderate dementia; and 3, for severe dementia. Between Stages 0 and 1 on the CDR is the Stage 0.5, ‘questionalble dementia’, which is described as mild memory loss with intact orientation, full achievement on the activities of daily living (ADL), with slight impairment in social activity but fully independent activity in daily life. Individuals in CDR Stage 0.5 show only slight forgetfulness, partial memory loss, with little impairment in time recognition and full recognition in space and person. Owing to the slightly impaired function of problem solving, these people may show slight impairment in social activity, but are usually able to live quite independent lives.

The GDS divides the span between the healthy brain to severe dementia into seven stages: Stage 1 is normal healthy, Stages 2–4 are prodromal stage of dementia, and Stages 5–7 are dementia. More precisely, GDS-1 is defined as the normal healthy stage, with full operation of all cognitive functions; GDS-2 represents the stage where subjective memory loss is noticed by subjects, but no objective memory dysfunction is revealed by clinical investigation; GDS-3 is the stage at which memory dysfunction is disclosed by clinical examination and this stage is named initially ‘mild cognitive decline’, which can be similar to ‘mild cognitive impairment’ (MCI), later proposed by Petersen et al.10 Subjects with GDS-3 may show slight impairment in social and professional activities, but full function in most of their daily life (Fig. 1).

Figure 1.

Scheme of ‘retrogenesis’ proposed by Reisberg and Gauthier.15 CDR, Clinical Dementia Rating Scale; FAST, Functional assessment of dementia of the Alzheimer type; MMSE, Mini-Mental State Examination; AD, Alzheimer's disease; SCI, Subjective cognitive impairment; MCI, Mild Cognitive Impairment.

Many concepts were proposed in the 1990s that describe the predementia stage with slightly different definitions. Examples include isolated memory impairment, incipient dementia, dementia prodrome, mild cognitive disorder, age-related cognitive decline, and mild neurocogntive disorder.

In the late 1990s, psychometric data were included to define the predementia stage in addition to clinical observations. Petersen et al. proposed MCI10 and Graham et al. proposed ‘cognitive impairment no dementia’.11 Originally, MCI was defined as: (i) subjective memory complaints; (ii) objective memory impairment; (iii) no cognitive impairment; (iv) full ADL; and (v) no dementia.10 Elderly people living in the community who fulfilled the MCI criteria (n = 66; mean age 81 years) were followed up to determine the rate of dementia 4 years later. The MCI subjects showed a rate of dementia of 12%/year, which is significantly higher than that of the general population.10 In European countries, the concept of ‘age-associated cognitive decline’ (AACD) was proposed,12 which is defined by more than one standard deviation (1 SD) below than the age-match general population in at least one of the five cognitive domains (memory, attention, spatial recognition, language, and speculation). The AACD takes into equal consideration the five domains of cognitive function, whereas the MCI regards memory as the main criterion. In one study,13 an elderly population living in the community was investigated; 19.3% of people matched the criteria given by the AACD, but only 3.2% matched the criteria of the MCI. After 3 years follow up, 28.6% of AACD subjects had developed dementia, compared with only 11.1% of MCI subjects. Based on these findings, the concept of the AACD is claimed to depict a more stable cohort of cognitively impaired elderly, and the AACD can more specifically select predementia subjects who are likely to develop dementia.13

Petersen et al. later developed the original MCI (amnestic MCI) into a broader concept of MCI.14 An amnestic MCI is an individual with memory impairment, and there may be multiple-domain MCI showing deficits in one of the five cognitive functions: memory, language, attention, executive function, and spatial recognition. In this way, the MCI can be subdivided into subtypes, multiple-domain MCI with and without amnesia, and single non-memory domain MCI. These subclassifications of MCI subjects may correspond to the prestage of Alzheimer's disease, Lewy body disease, frontotemporal dementia, vascular dementia, or depression, respectively14 (Fig. 2).

Figure 2.

Multiple domain mild cognitive impairment (MCI). AD, Alzheimer's disease; DLB, diffuse Lewy body disease; FTD, frontotemporal dementia; VaD, vascular dementia; Depr, depression. (Modified from Peterson et al.10).

However, the development of the MCI concept into subtypes is rather complicated and it can be better seen as the prestage of each corresponding disorder. In a sense, different subtypes of the MCI lose the significance of the concept of the MCI because it implies that MCI subjects are heterogeneous by nature, suggesting less usefulness of the conceptualization of the MCI as a whole.


The main criteria of the original MCI of Petersen et al.10 are subjective memory loss and objective memory dysfunction. The other three criteria are to exclude subjects with cognitive impairment other than memory, ADL disability, and dementia. Discussions have been raised as to whether there is any difference between subjective and objective memory loss, because these two criteria can be dependent on each other. More specifically, the question can be raised, what is the condition where subjective memory impairment is recognized without any evidence of objective memory dysfunction? This is an important question because it relates to the question whether subjective cognitive impairment (SCI) should be regarded as prestage of MCI.

Many people feel that their memory worsened after middle age. Many people admit that their memory worsened at 40–50 years of age, with people reporting having trouble recalling names of old friends or familiar places. Unfortunately, there is no scale or psychometric test to measure this type of subjective forgetfulness. The memory tests used routinely are not sensitive enough for this type of forgetfulness because they are designed to discriminate more severe cases of memory problems. We need more sensitive measures to detect fine decrements in memory function to show evidence related to impairment of subjective memory.

As described in literature, memory capacity is quite different between individuals, especially in the elderly. More differences in memory function are observed in older populations. What we need is a measure to estimate the reduction in memory function compared with the highest level of a given individual. The goal is so simple: we need to measure and compare the decrease in memory function of individual subjects from their highest level of memory function.

Dementia patients complain of memory loss only in the initial stages of the disease. The memory complaint will gradually disappear as the disease progresses. Alzheimer patients in earlier stage often recognize the fact that their memory function has been lost and complain of it. However, in the moderate to severe stages of Alzheimer's disease, most patients do not complain of memory loss because they lose the subjective recognition of memory function.

Conversely, patients with depression often complain about their memory function. Even though they feel subjectively strong impairment in memory function, they often show almost normal or little impaired memory on objective evaluation. Extreme cases have been observed, termed pseudodementia, who show significantly deteriorated ADL function with relatively preserved memory function. When such patients are examined for memory function, they show almost normal levels, but still complain of memory loss. The patients are good examples of cases in which subjective memory impairment does not parallel objective memory impairment. Subjective memory impairment is affected by emotion and mood. Then, we have to ask the question whether memory function or cognitive function, which can be affected by an individual's emotional state, can really be objective markers of cognitive function. A more important question is whether this type of subjective memory impairment can be a precedent for objective memory dysfunction. If SCI can be regarded as an earlier stage of MCI, we can draw a scheme of SCI, MCI, and dementia, as shown in Figure 3.

Figure 3.

Sequential scheme of subjective cognitive impairment (SCI), mild cognitive impairment (MCI), and dementia. ADAS.cog, Alzheimer's Disease Assessment Scale cognitive domain.


On the GDS staging, Reisberg and Gauthier15 differentiated GDS-2 and GDS-3 as two different prestages of dementia, the former with subjective memory complaints without objective memory dysfunction, and the latter with objective memory dysfunction. In this sense, GDS-2 corresponds to SCI and GDS-3 corresponds to MCI. As shown in Figure 3, Reisberg and Gauthier estimate the duration of MCI as 2 years and that of SCI as 15 years.15 Under the concept of retrogenesis,16 Reisberg et al. understand the process of functional changes in the brain after the highest level of functioning has been reached, most probably in people in their 30s.17

Human brain function is very primitive just after birth. We are not capable of memory, language, or communication, implying that most cognitive function is not yet equipped at birth. Even though most of the component material for network formation exists, neuronal networks are not yet established due to immature myelination of neuronal axons. In the human brain, myelination will be accomplished in interaction with environmental factors along with actual experience. The function of the neural networks is adjusted and fine-tuned for the best accomplishment through experience in actual life. It may take 15–30 years to reach the highest level of accomplishment. This process is controlled not only by biological factors, but also by psychosocial dynamics, in which parental bonding, human relationships, and social interaction will have a dominant influence on the functional level of the neuronal networks of a given individual. The highest level of cognitive functioning is quite different between individuals. There is a variety of behavioral patterns that is described as personality (i.e. temperaments or characteristics of individuals).

Then, after 30 years of age, the highest level of functioning will begin to deteriorate. As is postulated by Reisberg et al.,16 this downward process is retrogenetic because it goes back against the process of neural development, eventually leading to total disability of brain function. The long period of retrogenesis can be labeled as SCI.

Figure 4 shows the life-long variation in cognitive function, including SCI, MCI, and dementia stage. It is proposed that MCI corresponds to the stage where social life capacity is impaired with the full functioning of personal life capacity and biological life capacity. During the course of dementia, personal life capacity will deteriorate, but it is important to note that biological life capacity is well preserved, even in the late stage of dementia (Fig. 4).

Figure 4.

Life-long view of cognitive function. SCI, subjective cognitive impairment; MCI, mild cognitive impairment.


Even though the concept of SCI is immature and still has many problems that need to be resolved, the authors regard the concept of SCI useful. It should be refined further and the pathophysiology of SCI will attract more research attention, because earlier intervention in the pathogenesis of Alzheimer's disease is highly anticipated.

It is well described that amyloid deposition is observed 15–20 years before the onset of clinical symptoms. After the era of symptomatic pharmacotherapy, which ameliorates some symptoms of Alzheimer's disease, we are now aiming for disease-modifying therapy, and even its prevention. Early detection and prevention of Alzheimer's disease is a more important strategy for this disorder; therefore, the concept of MCI has been characterized and the methodology for the prevention of Alzheimer's disease has been studied.18–21 Among the three hallmarks of Alzheimer's disease of amyloid deposition, tau pathology, and neuronal degeneration, amyloid deposition is observed in the earliest stage, and the suppression of amyloid pathology is naturally the main target for drug development. The timing of amyloid deposition is certainly much earlier than the onset of clinical symptoms and even earlier than the onset of MCI. This is the why SCI should be the target of studies to develop therapies for the prevention of and early intervention in Alzheimer's disease. If SCI, MCI, and dementia can be arranged in this sequential order, of course the first pathological state should be the target of intervention. Even though it is not easy to develop a clear definition of SCI, clinical researchers will be expected to work more with subjects with SCI.