Gamma oscillations weaken with age in healthy elderly in human EEG

Abstract

Gamma rhythms (~20-70 ​Hz) are abnormal in mental disorders such as autism and schizophrenia in humans, and Alzheimer's disease (AD) models in rodents. However, the effect of normal aging on these oscillations is unknown, especially for elderly subjects in whom AD is most prevalent. In a first large-scale (236 subjects; 104 females) electroencephalogram (EEG) study on gamma oscillations in elderly subjects (aged 50-88 years), we presented full-screen visual Cartesian gratings that induced two distinct gamma oscillations (slow: 20-34 ​Hz and fast: 36-66 ​Hz). Power decreased with age for gamma, but not alpha (8-12 ​Hz). Reduction was more salient for fast gamma than slow. Center frequency also decreased with age for both gamma rhythms. The results were independent of microsaccades, pupillary reactivity to stimulus, and variations in power spectral density with age. Steady-state visual evoked potentials (SSVEPs) at 32 ​Hz also reduced with age. These results are crucial for developing gamma/SSVEP-based biomarkers of cognitive decline in elderly.

Keywords: Aging; Alpha oscillations; Alzheimer’s disease; EEG; Gamma oscillations; SSVEP.

Fig. 1. Slow and fast gamma in an example elderly subject.

a) Trialaveraged EEG trace (1st row, blue); time-frequency spectrograms of raw power (2nd row) and change in power from baseline (3rd row); and change in power with time (4th row) in alpha (8–12 Hz, violet), slow (20–34 Hz, pink) and fast gamma (36–66 Hz, orange) bands averaged across 10 unipolar (left column) and 9 bipolar (right column) electrodes. Vertical dashed lines represent actual stimulus duration (0–0.8 s, black) and period used for analysis within stimulus duration (0.25–0.75 s, red). Horizontal lines represent baseline (-0.5-0 s, black) and stimulus (0.25–0.75 s, red) analysis periods. White lines in spectrograms represent slow (solid) and fast (dashed) gamma frequency ranges. b) Right ordinate shows raw power spectral densities (PSDs, black traces) vs frequency in baseline (dotted) and stimulus (solid) periods averaged across 10 unipolar electrodes (left column) and 9 bipolar (right column) electrodes; left ordinate shows the same for change in PSD (in dB, solid blue trace) in stimulus period from baseline. Solid pink lines and dashed orange lines represent slow and fast gamma bands respectively. c) Scalp maps showing 112 bipolar electrodes (represented as disks). Color of each disk represents change in slow (left) and fast (right) gamma power. 9 electrodes used in 1a and 1b (right column) are marked with dots.

Fig. 2. Baseline PSDs, slopes and alpha power.

Baseline PSDs (averaged across 10 unipolar or 9 bipolar electrodes) for three age-groups on a log-log scale for unipolar (left) and bipolar (right) reference, plotted for males (2a) and females (2b). Thickness of traces indicate SEM across subjects. Age-group limits and the number of subjects in the respective age-groups are indicated on the left plot. c) Same as in 2a and 2b, but for males and females, pooled across all age-groups. d) Mean baseline PSDs for three ranges of baseline absolute alpha power (8–12 Hz, power ranges for respective traces indicated on the plots) pooled across all age-groups. Thickness of traces and numbers indicate SEM across subjects and number of subjects in respective alpha power ranges. Colored bars on the abscissa indicate alpha (8–12 Hz, violet), slow (20–34 Hz, pink) and fast gamma (36–66 Hz, orange) frequency bands.

Fig. 3. Slow and fast gamma in younger and elderly subjects.

a) Scatter plot showing change in slow (abscissa) and fast (ordinate) gamma power. Dotted lines represent 0.5 dB threshold. Points represent subjects with no gamma (dark blue), only slow gamma (light blue), only fast gamma (green) and both gamma rhythms (yellow) with change in power above 0.5 dB threshold. b) Change in PSDs vs frequency averaged across subjects (numbers denoted by n) as categorized in 3a. Thickness of traces indicate SEM. Solid pink and dashed lines represent slow and fast gamma ranges respectively. c) Bar plot showing percentage of subjects in three age-groups (marked by respective colors) categorized as in 3a. d) Schematic showing placements of left and right anterolateral and posteromedial group of bipolar electrodes used for analysis on the scalp, as well as ground (Gnd) and online reference (Ref) electrodes. e) Average scalp maps of 112 bipolar electrodes (disks) for three age-groups for slow (top row) and fast (bottom row) gamma. Color of disks represents change in respective gamma power. Electrode groups represented as in 3d.

Fig. 4. Change in gamma power vs age for anterolateral group of electrodes.

Mean time-frequency change in power spectrograms (4a) and change in power spectra vs frequency (4b) for three age-groups separately for males (top row) and females (bottom row). Thickness of traces and numbers in 4b indicate SEM and number of subjects respectively. Solid and dashed lines indicate slow and fast gamma frequency ranges respectively. c) Left column: bar plots showing mean change in slow gamma power for three age-groups separately for males and females. Number of subjects for respective age-groups are indicated on top. Error bars indicate SEM. Right column: scatter plot for change in slow gamma power vs age for all elderly subjects (>49 years age-group, n = 227), plotted separately for males (in orange) and females (in yellow). Orange, yellow and black solid lines indicate regression fits for males, females and data pooled across gender respectively. p-values of the regression fits are indicated in respective colors. d) Same as in 4c but for fast gamma.

Fig. 5. Center frequency of slow and fast gamma vs age for elderly subjects for anterolateral group of electrodes.

Scatter plots showing center frequency vs age for slow and fast gamma, for anterolateral electrodes, for those subjects who have change in power in respective gamma range above 0.5 dB (numbers indicated on the plots). Solid lines indicate regression fits for center frequency vs age. p-values for these fits are as indicated.

Fig. 6. Change in alpha power vs age.

a) Left column: bar plots showing mean change in alpha power across anterolateral group of electrodes for three age-groups separately for males and females. Number of subjects for respective age-groups are indicated at bottom. Right column: scatter plot for change in alpha power vs age for all elderly subjects (>49 years age-group, n = 227), plotted separately for males (in orange) and females (in yellow). Same format as in Fig. 4c b) Scalp maps for 112 electrodes (disks) averaged across all subjects separately for three age-groups. Color indicates change in alpha power for each electrode, same format as in Fig. 3e.

Fig. 7. Eye position, microsaccades and pupillary reactivity across age for elderly subjects.

a) Eye-position in horizontal (top row) and vertical (middle row) directions; and histogram showing microsaccade rate (bottom row) vs time (-0.5–0.75 s of stimulus onset) for elderly subjects (n = 226). Number of subjects in each age-group is indicated on top. Thickness indicates SEM. b) Main sequence showing peak velocity and maximum displacement of all microsaccades (number indicated by n) extracted for both elderly age-groups. Average microsaccade rate (mean ± SEM) across all subjects for each elderly age-group is also indicated. c) Scatter plot showing change in power vs age for slow (top row) and fast (bottom row) gamma for all elderly subjects with analyzable data after removal of trials containing microsaccades. Solid lines indicate regression fits. Numbers of subjects with analyzable data in each age-group is indicated on top. d) Scatter plots for coefficient of variation (CV) of pupil diameter vs age (top row), change in slow (middle row) and fast (bottom row) gamma power. Pearson correlation coefficients (r) and p-values are also indicated.

Fig. 8. Change in SSVEP power vs age for anterolateral group of electrodes.

Time-frequency change in power spectrograms (8a) and change in power spectra vs frequency (8b) for three age-groups separately for males (top row) and females (bottom row). Thickness of traces in 8b indicates SEM. Insets in 8b display zoomed-in images of respective main plots, showing clear SSVEP peaks at 32 Hz. c) Top row: bar plots showing mean change in SSVEP power for three age-groups separately for males and females; numbers of subjects in each age-group is indicated on top. Error bars indicate SEM. Bottom row: scatter plot for change in SSVEP power vs age for all elderly subjects (>49 years age-group, n = 197), plotted separately for males (in orange) and females (in yellow). Orange, yellow and black solid lines indicate regression fits for males, females and data pooled across gender respectively. p-values of the regression fits are indicated in respective colors. d) Scalp maps for 112 electrodes (disks) averaged across all subjects separately for three age-groups. Color indicates change in SSVEP power at 32 Hz for each electrode.