Abstract
In both humans and dogs sleep spindle occurrence between acquisition and recall of a specific memory correlate with learning performance. However, it is not known whether sleep spindle characteristics are also linked to performance beyond the span of a day, except in regard to general mental ability in humans. Such a relationship is likely, as both memory and spindle expression decline with age in both species (in dogs specifically the density and amplitude of slow spindles). We investigated if spindle amplitude, density (spindles/minute) and/or frequency (waves/second) correlate with performance on a short-term memory and a reversal-learning task in old dogs (> 7 years), when measurements of behavior and EEG were on average a month apart. Higher frequencies of fast (≥ 13 Hz) spindles on the frontal and central midline electrodes, and of slow spindles (≤ 13 Hz) on the central midline electrode were linked to worse performance on a reversal-learning task. The present findings suggest a role for spindle frequency as a biomarker of cognitive aging across species: Changes in spindle frequency are associated with dementia risk and onset in humans and declining learning performance in the dog.
Introduction
Sleep spindles are brief trains of rhythmic activity, at least half a second in duration and maximally 6 seconds long, which appear in the EEG signal of humans, and other mammals during non-REM sleep, in particular stage 2 of non-REM sleep in humans. They are commonly distinguished in a slow (predominantly frontal, ≤ 13 Hz) and fast (≥ 13 Hz, predominantly central and posterior) subtype.
One promising model animal in comparative sleep spindle research is the dog (Canis familiaris). A shared anthropogenic environment and evolutionary adaptation to its dynamics characterize dogs as an animal model for human conditions in general, and they have specifically been argued a favorable model in comparative neuroscience as well. Moreover, there is recent evidence, that, in dogs, transients oscillating in the 9–16 Hz frequency range, corresponding to the broad definition of the sigma band or spindling frequency in humans, are analogous to human sleep spindles (See Table for an overview of these analogies and the associated literature).
As a note of caution, regarding spindle-cognition associations in general, detection methods for sleep spindles tend to correlate poorly with each other, which is challenging to the comparability between studies. In addition, sleep spindles also appear to display a link to memory consolidation and learning only under specific circumstances. The distance of the to-be-learned material to prior knowledge, the exact stage of non-REM sleep in which spindles are measuredand the timing relative to cortical up-states and ripples, all seem to have an influence. In light of this it is not surprising that associations between sleep spindles and learning are not always replicated. Considering the controversy about the comparability of different spindle detection methods, it is important to note that both studies to date on spindle expression in dogs (Iotchev et al.) used the same search criteria. The two studies are therefore directly comparable with each other and the present study. These dog spindle investigations have replicated several findings in the human literature, although dog sleep spindles also display some unique dynamics. In particular frontal, fast (≥ 13 Hz) spindles do not decline, but increase with age in the dog.
Age seems to affect both memory and the expression of sleep spindles in humans and dogs, but it has seldom been investigated in either species how changes across time in one variable associates with changes in the other. Attempts thus far have either failed to show this type of association or entirely omitted difference score analyses. Even when spindle activity (e.g. density, amplitude) was recorded at different time points or across age-groups it was only compared with immediate post-sleep behavior (the examples include research on early development and IQ). A third common line of research has revealed associations between changes in spindle characteristics and the risk for developing dementia, however the syndrome was not broken down into specific domains (e.g. memory or problem solving decline) so that it is not known which functions specifically change as spindle expression changes.
Day-to-day fluctuations in spindle occurrence and frequency may explain why difference score comparisons have mostly been avoided. Sources of such variation include the menstrual cycle and simple exposure to novel information, which increases spindle occurrence in humans, rats, and dogs. It seems, therefore, that without controlling for hormonal levels and pre-sleep experience, comparing spindle activity and memory performance measured further away than a day from each other could be problematic. On the other hand, relatively stable spindle occurrence has been reported within individuals, across nights, and the frequency ranges of spindling are also assumed to be relatively stable within individuals. Furthermore, age-related decline of spindle occurrence and amplitude in humans are a convincingly replicated effect which is not masked by the expected fluctuation.
In conclusion, differences in spindle characteristics measured between widely spaced time intervals could be useful biomarkers as they more likely reflect developmental or age-related changes, but more studies need to undertake difference score comparisons to test this assumption. Moreover, although useful for studying memory consolidation, same-day correlates of memory and learning are unlikely to reflect the general performance level of the individual. This is because of the already discussed sources of day-to-day variation in, for instance, occurrence, but also frequency of spindles. Same-day correlates between sleep spindles and learning are thus of low diagnostic value. It has been established that spindle amplitudes and duration in humans correlate with IQ even when each is measured more than a day apart from the other. No work has investigated the same for learning and memory, and the effect sizes for IQ are also modest, though well-replicated. From a clinical point of view, it is a very important question concerning possible applications of diagnostic EEG in the veterinary praxis if a measurement can predict performance further apart in time.
In the present study we apply the same automatic spindle detection algorithm as previously used in the dog to a data-set containing two measurements per dog, of each EEG and behavior (a short-term memory task and a reversal-learning paradigm) to evaluate whether canine spindles are useful markers of cognitive aging. First, we investigated if changes between two widely spaced (at least 3 months apart) samples of EEG can predict changes in performance between two similarly timed measures of the two learning tasks. Second, because the interval between corresponding probes of EEG and behavior vary from several days to about a month (on average) the data also allowed us to test if any spindle features can function as markers beyond the span of a day. Finally, we focused on a sample of exclusively older dogs (minimum age = 7 years), as variation in performance in elderly individuals is likely better suited to observe markers relevant for distinguishing healthy and pathological aging. The sample from which the present selection of dogs was taken (based on the availability of EEG data) was demonstrated before to display a variability in cognitive performance.
We hypothesized that occurrence, measured as density (spindles/minute), and/or amplitude would show a positive association with performance on the two tasks, while we expected an inverse relationship with spindle frequency, as it rises with age in humans and dogs. Importantly, as done in many human studies, we separately analyzed fast and slow spindles as in both humans and dogs age-related changes in spindle characteristics differ between the slow and fast variety and topographically (for instance more pronounced slow spindle amplitude decline over the frontal cortex in both humans and dogs).
Data: 16 aprile 2020
Authors: Ivaylo Borislavov Iotchev, Dóra Szabó, Anna Kis, Enikö Kubinyi.
Fonte: www.nature.com