electroencephalography - aged, 80 and over - health of the elderly - aging
eletroencefalografia - idoso de 80 anos ou mais - saúde do idoso - envelhecimento
Brain tissue has an enormous quantity of connections between neurons. These become
established so that neuronal activities become synchronous and different brain areas
can communicate. One of the mathematical parameters that enable measurement of this
connectivity is electroencephalographic (EEG) coherence. Several published studies
have suggested that this coherence decreases through the aging process, with reduction
of the neuron population[1].
Life expectancy in Brazil has increased over recent years and, consequently, the population
of nonagenarians has increased. Therefore, there is a need for knowledge of the neurophysiological
characteristics of EEG signals in cognitively normal individuals of this age.
Investigation of EEG coherence enables measurement of the correlation between different
brain areas. This is done through examining the mutual relationship between two EEG
signals for different frequency bands. Quadratic or ordinary coherence consists of
measurement of the covariance of spectral power between two pairs of electrodes[1],[2],[3]. Because coherence is a normative value for the power of a given frequency, it is
independent of the amplitude oscillations of the signal[4],[5]. Thus, this property makes coherence a relatively safe method for comparing groups
of individuals, because differences in EEG potency do not affect the estimated coherence
value[6]. Diminished coherence between two regions probably indicates a decrease in their
functional correlations[1],[7].
It is possible to calculate the ordinary coherence between two brain electrodes that
are placed in the same hemisphere, or in homologous areas, through a mathematical
formula that divides the cross-spectrum of these two channels by the product of the
auto-spectrum of each of the same channels, as demonstrated by the formula below:
Co is the coherence function
Gxx and Gyy are the spectral power of each channel
Gxy is the power of the cross-spectrum
The coherence value between two electrodes varies from 0 to 1, where 1 means the maximum
correlation, and 0 means that the correlation is very small or absent[5]. Other studies, such as that by Koyama et al.[8], who evaluated 68 normal elderly people (61−90 years old) compared to 20 young people
(23−50 years), showed that age had no effect on interhemispheric coherence, although
intrahemispheric coherence was reduced in all bands, almost linearly.
Anghinah et al.[2] did not find any significant differences in interhemispheric coherence for the alpha
band between individuals in older age groups and young adults. Jorge et al.[3] studied a population of normal adults and did not observe any influence from age,
gender or manual dominance regarding interhemispheric coherence values, for any of
the frequency bands evaluated.
Thus, the objective of the present study was to analyze interhemispheric coherence
between O1O2 electrodes in a population of cognitively normal nonagenarians.
METHODS
In this study, 80 individuals aged 90 years or older were evaluated. These individuals
were recruited from the community or were among a group of elderly patients followed
up at the outpatient clinic of the Cognitive and Behavioral Neurology Group and the
Reference Center for Cognitive Disorders of the Neurological Clinical Division of
Hospital das Clínicas, University of São Paulo Medical School, in São Paulo, or were
from sheltered housing for elderly people in São José dos Campos.
Out of this total, 42 participants, who were cognitively normal according to the Mayo
Clinic’s Older Americans Normative Studies 1992 criteria[9], were included (52.5%). These participants did not present with any complaints regarding
memory and all of them were living independently. None of them had any disease that
could compromise their cognitive functions and none of them were using any medication
with psychotropic effects or any effects on cognition.
They underwent an interview, a neurological clinical examination, a neuropsychological
evaluation and a digital EEG.
The cognitive assessment followed a protocol that included the Mini-Mental State Examination
(MMSE)[10], Pfeffer’s Functional Activities Questionnaire[11], Katz assessment[12], Dementia Rating Scale[13], verbal fluency evaluation (semantic and phonemic) and digit repetition assessment
in direct order and inverse order.
All the subjects underwent electroencephalography in a silent environment, with closed
eyes, after having been at rest for a period of 30 minutes. The recording used digital
low-pass and high-pass filters with cutoff frequencies at 0.5 and 30 Hz.
The electrode placement system followed the 10-20 standard as recommended by the Brazilian
Society of Clinical Neurophysiology[14],[15]. The electrode impedance was less than 5.0 ohms and a technically-linked binaural
reference was used, which is the most appropriate form for this type of recording[16]-[18].
From these recordings, 20 records of 2.56 seconds each were selected. These were free
from artifacts, such as eye movements, eyelid movements and muscle potentials, and
were free from transient findings such as benign variants[19]. The fast Fourier transform was applied to obtain spectral analyses with the respective
values for each region of the scalp. After obtaining the spectra, the interhemispheric
coherence was calculated for all the bands. The equipment used was an EMSA Braintech
digital EEG system with 32 channels, 200 Hz sampling frequency and 12-bit processor,
with software capable of performing quantitative evaluations and coherence analyses
through application of the fast Fourier transform.
All the participants agreed to undergo the protocol by signing a free and informed
consent statement, which was done either by the individuals themselves or by a member
of their family. This project was approved by the Ethics Committee for Analysis of
Research Projects (CAPPesp), which is part of the Clinical Board of Hospital das Clínicas,
University of São Paulo Medical School, under number 0500/10.
RESULTS
Among the 80 individuals who underwent neuropsychological assessment, 33 were excluded
because they presented with mild cognitive impairment or even the beginnings of dementia.
Five of the potential participants dropped out, i.e. they decided that they did not
want to complete the evaluation, for a variety of reasons. The remaining 42 individuals
met the inclusion criteria. In this study, the ages of the participants ranged from
90 to 101 years, with an average of 92 (SD 3), with the largest proportion of them
being 90 years of age (39%) ([Table 1]).
Table 1
Sample characteristics.
|
Variable
|
n (%)
|
|
Age (years)
|
|
90−95 years
|
38 (90.47)
|
|
95−100 years
|
3 (7.14)
|
|
> 100 years
|
1 (2.38)
|
|
Gender
|
|
Men
|
12 (29)
|
|
Women
|
30 (71)
|
|
Schooling level
|
|
0−5 years
|
12 (28.60)
|
|
6−10 years
|
10 (23.80)
|
|
11−15years
|
15 (35.70)
|
|
> 15years
|
5 (11.90)
|
|
Total
|
42 (100)
|
The educational level ranged from 0 (illiterate) up to 19 years, with an average of
nine years (SD 4.60). Regarding the gender of the individuals studied, 29% (12) were
male and 71% (30) were female. The MMSE scores ranged from 20 to 30, and the mean
was 27.14 (SD 2.60). The individual with the lowest MMSE score (20) was the one with
the lowest level of education. The largest proportion of the participants (10 individuals;
23.80%) had an MMSE score of 29.
The Pfeffer Functional Activities Questionnaire score[11] ranged from 0 to 13, with an average of 1.85 (SD 3.0). The highest score was for
an individual with motor and visual difficulties due to bilateral hip arthrosis and
diabetic retinopathy. Out of the total group, 23 individuals (54.76%) had Pfeffer
scores of 0.
The Katz index ranged from 5 to 6, with an average of 5.90 (SD 0.30). The largest
proportion of the participants (38 individuals; 90%) had a Katz index of 6.
The Dementia Rating Scale ranged from 112 to 139, with a mean of 123 (SD 12.5). Semantic
verbal fluency (in which the participants were asked to list supermarket items) ranged
from 9 to 25, with a mean of 16 (SD 4.0).
The repetition of digits in direct order ranged from 4 to 6, with an average of 4.73
(SD 0.91) and for inverse order the score ranged from 2 to 6, with a mean of 3.52
(SD 0.91).
The analysis of interhemispheric coherence for the alpha band at O1O2 showed an average
value of 0.65 (SD 0.13), and the value that appeared most frequently was 0.65 (14.63%
- median), with a range of 0.48 to 0.98.
DISCUSSION
We analyzed, in our study in 2007[3], the interhemispheric coherence for the alpha band for eight pairs of electrodes
distributed across the scalp, in normal individuals aged 20 to 50 years. The mean
values found were 0.66 for P3 and P4, and 0.63 for O1 and O2. Thus, we observed that
there was no significant difference between them (Mann-Whitney U test for independent
samples with p < 0.05) ([Figure]).
Figure Alpha bandO1O2 coherence frequency between age groups.
However, in a study by Duffy et al.[20], with 37 individuals aged 20−80 years, a significant decrease in coherence was found
with increasing age. Moreover, in a study by Kikuchi et al.[21], on interhemispheric coherence for several frequency bands among 55 healthy volunteers,
comparing young and old, decreased coherence among older individuals (50−67 years)
was also found.
Normal aging in humans is known to affect several aspects of central cholinergic function.
This results in deficits in acetylcholine release levels and changes to muscarinic
receptor plasticity. Hence, Sloan et al.[22] demonstrated that when scopolamine (an anticholinergic drug) was administered to
normal subjects, it decreased interhemispheric coherence in the alpha and beta bands.
These findings suggest that a reduction in cholinergic activity is involved in the
underlying normal mechanism of reduced coherence in elderly individuals.
Nevertheless, several different mechanisms appear to be involved in modulation of
EEG coherence, such as corpus callosum integrity and thalamo-cortico-thalamic reverberation.
Koeda et al.[23] and Montplaisir et al.[24] demonstrated the importance of the corpus callosum in this analysis. These authors
showed that individuals with agenesis of the corpus callosum, or those who were subjected
to partial corpus callosotomy, presented with reduced interhemispheric coherence.
In our study, the maintenance of interhemispheric coherence values may have been due
to the integrity of the corpus callosum in this age group. On the other hand, Weiss
et al.[25] demonstrated, through MRI, a reduction of the corpus callosum in normal elderly
individuals. However, their participants were younger than ours. Moreover, among our
individuals who underwent an MRI, no abnormality affecting the corpus callosum was
detected.
Hogan et al.[26] studied the effects of aging on EEG coherence through comparison between young adults
and normal elderly individuals. They also found that the coherence was lower among
their healthy elderly group. However, they compared intrahemispheric coherence between
the frontotemporal, frontocentral and temporoparietal regions, unlike our study in
which we evaluated interhemispheric coherence. Other studies, such as that of Vysata
et al.[18], demonstrated declines in the magnitude of interhemispheric coherence for the alpha
and theta bands with aging, caused by reductions in cerebral connectivity.
Another factor that can interfere in the calculation of coherence between two points
distant from the scalp is the type of reference used during the recording. Therefore,
we used a linked binaural reference, which, according to data in the literature, such
as in the study by Vysata et al.[18], is the technique that interfered least in determining coherence in relation to
the age of the individuals studied.
In the present study, we also compared interhemispheric coherence for the alpha band
between men and women, and between the P3P4 and O1O2 pairs of electrodes. We did not
find any statistically significant differences (Mann-Whitney U test; p < 0.05). This
may have been due to the small size of our sample, the prevalence of women (3:1),
or the age group studied.
Our results were very close to those obtained by Anghinah et al.[2],[27] in Brazilian populations over the age of 50 years ([Table 2]).
Table 2
Comparison between several studies on alpha band coherence at O1O2.
|
Author
|
Age
|
Coherence mean
|
|
Anghinah, 200027
|
> 50 years
|
0.658
|
|
Anghinah, 20052
|
> 50 years
|
0.660
|
|
Jorge, 20073
|
< 50 years
|
0.637
|
|
Jorge, 2017
|
90−100 years
|
0.657
|
In addition, we did not find any correlation between MMSE scores and interhemispheric
coherence values for the alpha band at O1O2 (Pearson’s correlation coefficient 0.31;
p = 0.05).
In conclusion, the results obtained showed that the mean coherence value for the alpha
band in the occipital regions in normal nonagenarian individuals was 0.65, with a
range from 0.48 to 0.98. Therefore, our study did not show any decrease in interhemispheric
coherence for the alpha band at the O1O2 electrodes in this group. This may indicate
that this represents a standard value for this age group. We also believe that further
studies with a larger number of participants and with imaging tests may provide greater
knowledge on this subject.
