ORIGINAL RESEARCH
Differences in EEG Delta Frequency Characteristics and Patterns in Slow-Wave Sleep Between Dementia Patients and Controls: A Pilot Study Enrica Bonanni,* Elisa Di Coscio,* Michelangelo Maestri,* Luca Carnicelli,* Hara Tsekou,† Nicholas Tiberio Economou,† Thomas Paparrigopoulos,† Anastasios Bonakis,‡ Sokratis G. Papageorgiou,‡ Dimitris Vassilopoulos,‡ Constantin R. Soldatos,† Luigi Murri,* and Periklis Y. Ktonas†
Purpose: To evaluate the modifications of EEG activity during slow-wave sleep in patients with dementia compared with healthy elderly subjects, using spectral analysis and period-amplitude analysis. Methods: Five patients with dementia and 5 elderly control subjects underwent night polysomnographic recordings. For each of the first three nonrapid eye movement–rapid eye movement sleep cycles, a well-defined slow-wave sleep portion was chosen. The delta frequency band (0.4–3.6 Hz) in these portions was analyzed with both spectral analysis and period-amplitude analysis. Results: Spectral analysis showed an increase in the delta band power in the dementia group, with a decrease across the night observed only in the control group. For the dementia group, period-amplitude analysis showed a decrease in well-defined delta waves of frequency lower than 1.6 Hz and an increase in such waves of frequency higher than 2 Hz, in incidence and amplitude. Conclusions: Our study showed (1) a loss of the dynamics of delta band power across the night sleep, in dementia, and (2) a different distribution of delta waves during slow-wave sleep in dementia compared with control subjects. This kind of computer-based analysis can highlight the presence of a pathologic delta activity during slow-wave sleep in dementia and may support the hypothesis of a dynamic interaction between sleep alteration and cognitive decline. Key Words: Sleep, Slow-wave sleep, Period-amplitude analysis, Spectral analysis, Dementia. (J Clin Neurophysiol 2012;29: 50–54)
S
leep disturbances are very common in dementia patients and often lead to institutionalization because these disturbances are very stressful to caregivers (Ancoli–Israel and Vitiello, 2006; Bliwise, 2004). Sleep in dementia is characterized by an altered regulation of sleep–wake cycles, with an increased nocturnal sleep fragmentation and daytime sleep (Vitiello et al., 1992), probably related to a disruption of the circadian system and an inability of generating nonrapid eye movement (NREM) sleep (Wu and Swaab, 2007). As far as sleep organization in dementia is concerned, a decrease in total sleep time and rapid eye movement (REM) sleep time and an increase in arousals and awakenings have been observed, such modifications paralleling cognitive decline (Moe et al., 1995; Vitiello From the *Department of Neurosciences, University of Pisa, Pisa, Italy; †Sleep Research Unit, Department of Psychiatry, University of Athens, Athens, Greece; and ‡Cognitive Neurology-Extrapyramidal Disorders Unit, Department of Neurology, University of Athens, Athens, Greece. Supported in part by EU FP6 Project BIOPATTERN. Address correspondence and reprint requests to Enrica Bonanni, Department of Neurosciences, University of Pisa, Pisa, Italy; e-mail
[email protected]. Copyright Ó 2012 by the American Clinical Neurophysiology Society
ISSN: 0736-0258/12/2901-0050
50
et al., 1992). For slow-wave sleep (SWS), conflicting results have been observed, including a decrease, an increase, or no variation thereof (Bliwise, 1993). Two different aspects should be underlined. First, it is debated whether SWS declines with age because slow EEG waves (e.g., delta waves) have exhibited a reduction in amplitude but not in incidence and a deficiency in homeostatic reactivity (Darchia et al., 2007; Ohayon et al., 2004). Second, it should be emphasized that an increase in slow EEG frequencies (delta and theta) and a decrease in fast ones (alpha and beta) is commonly seen in dementia patients during wake (Babiloni et al., 2004; Jeong, 2004). This slowing of the EEG could continue during sleep, leading to a difficulty in differentiating pathologic from normal delta EEG activity (Crowley et al., 2005). Recent observations and hypotheses have highlighted the role of SWS and EEG slow-wave activity in cognitive processes (Tononi and Cirelli, 2006), as well as the fact that slow-wave activity is inhomogeneous concerning the sleep function and regulation attributed to its various frequency components (Darchia et al., 2007; Feinberg et al., 1987; Steriade et al., 2001). The aim of the present pilot study, which involves a relatively small number of subjects, was to investigate if computer-based EEG processing using spectral analysis (SA) and period-amplitude analysis (PAA) methodologies could distinguish differences in delta EEG activity during night SWS between dementia patients and healthy elderly subjects.
METHODS Subjects and Polysomnographic Evaluation
Five patients with dementia (mean age 6 SD: 69.8 6 4.1 years; 3 women and 2 men) and 5 healthy control subjects, matched for age and sex (mean age 6 SD: 70.2 6 4.4 years; 3 women and 2 men), were enrolled in the study. The dementia patients were inpatients at the Cognitive Neurology-Extrapyramidal Disorders Unit of the University of Athens, who were diagnosed for dementia through standard clinical, neuroimaging, and neuropsychologic procedures (e.g., DSM IV criteria). The patients had cortical dementias, including early Alzheimer disease, moderate frontotemporal dementia, posterior cortical atrophy, and frontotemporal dementia/semantic dementia. Both patients and control subjects were drug naive and were not taking any drugs during the sleep recordings. The protocol was approved by appropriate authorities, and an informed consent was obtained from patients and control subjects. Night polysomnographic data were obtained with standard procedures, for both patients and control subjects, using a Brain Quick (Micromed, Mogliano Veneto, Treviso, Italy) digital data acquisition system, using a sampling frequency of 128 Hz, at 16-bit resolution, and
Journal of Clinical Neurophysiology Volume 29, Number 1, February 2012
Journal of Clinical Neurophysiology Volume 29, Number 1, February 2012
using an EEG analog filter lower cutoff at 0.15 Hz. The EEG electrodes (F3, F4, C3, C4, P3, P4, O1, and O2) were positioned according to the 10–20 international system, and the electrode impedance was less than 5 kU. In addition to the EEG, electrooculogram and submental electromyogram were recorded for the detection of sleep stages. Visual sleep scoring was performed according to standard criteria, using 30-second epochs (Rechtschaffen and Kales, 1968). For every subject, based on the hypnogram, a NREM/REM sleep cycle was chosen in the beginning, middle, and end of the recording. Well-defined SWS periods, 1 per cycle, lasting 5 minutes each, named SWS1, SWS2, and SWS3, were visually selected for analysis. The SWS periods were chosen by two sleep experts, and any discordance was discussed and settled between them. Every SWS period was subdivided into consecutive 30-second epochs. Spectral analysis and PAA of the C3-G2 (G2 at the forehead) EEG derivation was performed for each 30-second epoch.
Spectral Analysis The EEG power spectrum (absolute values) was calculated via fast Fourier transform procedures for every 5-second data subepoch, with the Analysis Manager-Rembrandt (Micromed) software. The total delta EEG frequency band (0.4–3.6 Hz) as well as 8 delta EEG frequency subbands from 0.4 Hz to 3.6 Hz with a frequency resolution of 0.4 Hz each (i.e., d1:0.4–0.8 Hz; d2: 0.8–1.2 Hz; d3: 1.2–1.6 Hz; d4: 1.6–2.0 Hz; d5: 2.0–2.4 Hz; d6: 2.4–2.8 Hz; d7: 2.8– 3.2 Hz; d8: 3.2–3.6 Hz) were considered. The power spectrum for a 30-second data epoch was calculated as the average of 6 successive 5-second data subepoch spectra.
Period-Amplitude Analysis Every 30-second epoch was prefiltered between 0.008 and 20 Hz and underwent PAA in the delta EEG frequency band (Ktonas and Gosalia, 1981) with a computer-aided method (System Plus; Micromed). Visually well-defined delta waves were selected for analysis. A wave was heuristically defined as the EEG signal between two consecutive local voltage minima above the background “noise” level. For each wave, the wave frequency was calculated as the inverse of the wave period (measured time interval between the two consecutive voltage minima of the wave), and the wave amplitude was calculated as the voltage difference between wave minimum and wave maximum. Percentage of incidence for delta waves and mean amplitude of delta waves in each of the delta frequency subbands were calculated for each 30-second epoch.
EEG Delta Characteristics in SWS in Dementia
we compared the value of (mean 1 2 SD) for the control subjects versus every single case of dementia.
RESULTS Concerning the sleep macrostructure, there were differences in sleep variables between patients and control subjects, indicating the possibility of an alteration of sleep structure for the patients (Table 1). However, these differences were not statistically significant. Comparing the dementia and control groups for the sum of the SWS periods analyzed, SA showed a significant increase in the total power of the delta frequency band and in the power of all the delta frequency subbands, except for the first one (0.4–0.8 Hz), for the dementia group (Fig. 1). Considering NREM/REM sleep cycle effects in the control group, a significant power decrease for the whole delta frequency band and for all the delta frequency subbands, except the first one (0.4–0.8 Hz), was observed between SWS1 and SWS2 (Fig. 2). A tendency toward a further decrease was observed between SWS2 and SWS3. The dementia group showed a loss of these dynamics for the delta power during sleep (Fig. 3). Comparing the dementia and control groups for the sum of the SWS periods analyzed, PAA showed a significant decrease in the percentage of incidence and in mean amplitude of visually welldefined delta waves with frequency lower than 1.6 Hz, for the dementia group. However, it showed a significant increase in the percentage of incidence and in mean amplitude of visually welldefined delta waves with frequency higher than 2.0 Hz for the dementia group (Figs. 4 and 5). The analysis of the data from each of the dementia patients compared with the summed data of the control subjects showed similar differences of the same statistical significance as before (data not shown). Moreover, concerning the PAA results, the values of most patients were higher than the value of (mean 1 2 SD) of the control subjects for frequencies higher than 2.4 Hz (Tables 2 and 3).
DISCUSSION General Comments The computer-based quantification of SWS described in this work found differences in EEG delta frequency characteristics and patterns between dementia patients and control subjects. Dementia patients exhibited a reduction in both incidence and amplitude for slow delta waves (0.4–1.6 Hz), while they exhibited an increase in
Statistical Analysis Differences for total power in the delta EEG frequency band and for power in the delta EEG frequency subbands (obtained from SA) and differences for percentage of incidence and for mean amplitude in the delta EEG frequency subbands (obtained from PAA) were analyzed with analysis of variance (ANOVA), considering group (patients vs. control subjects) as the independent variable. The kinetics of the spectral power variables across the three NREM/REM sleep cycles were analyzed with ANOVA for repeated measures; ANOVAs were followed by post hoc t-tests. Because the 8 frequency subbands may not be considered as independent, the possible effect on the degrees of freedom was taken into account by considering significance at 0.01 for all ANOVA and t-test results. To investigate the variability among the patients, we performed ANOVA and t-tests for every single case of dementia versus the summed data for the group of controls. Moreover, for the PAA results, Copyright Ó 2012 by the American Clinical Neurophysiology Society
TABLE 1. Sleep Variables in Control Subjects and Dementia Patients (n ¼ 5) Control Dementia Subjects (n ¼ 5) Patients (n ¼ 5) Time in bed (minutes) Total sleep time (minutes) Sleep latency (minutes) Sleep efficiency (%) Wake after sleep onset (minutes) Stage 1 (%) Stage 2 (%) SWS (%) REM sleep (%) REM latency (minutes)
512.7 385.1 37.6 75.3 90.0 12.4 49.8 21.8 15.9 94.2
6 6 6 6 6 6 6 6 6 6
30.6 56.3 29.4 12.2 52.6 5.9 12.7 12.1 5.4 14.4
482.9 319.7 43.1 66.4 120.1 13.5 39.5 28.5 18.4 99.4
6 6 6 6 6 6 6 6 6 6
34.0 24.6 35.8 5.9 48.4 14.8 17.1 12.3 9.8 12.7
51
E. Bonanni et al.
Journal of Clinical Neurophysiology Volume 29, Number 1, February 2012
FIG. 1. Spectral analysis for the sum of all the SWS periods: comparison of absolute delta power (total and in subbands) between patients and control subjects. *P , 0.0001, **P , 0.0038, #P , 0.0011, and §P , 0.0007. both incidence and amplitude for fast delta waves (2.0–3.6 Hz). Furthermore, the dementia patients exhibited more total power in the delta frequency band, while they apparently exhibited a disturbed delta-related homeostatic mechanism across NREM/REM sleep cycles during the night sleep. The results indicate that SWS may be quantitatively altered in dementia as far as its electrographic structure and regulation is concerned. It should be noticed that our patients had various subtypes of the cortical dementias. Thus, the results are not specific to one type of abnormality but may relate to cortical impairment. Nevertheless, the comparison of individual patient cases versus the summed data of control subjects indicated significant differences between the individual cases and the control group. However, given the relatively small number of subjects in this pilot study, the findings should be considered as preliminary, and more studies are needed to examine these differences further.
Methodological Issues Comparing the results obtained from the two analysis methods, a concordance is seen for delta frequency subbands higher than 2 Hz, while a discrepancy is observed for slower delta waves. Because power is affected both by wave amplitude and by wave incidence, the fact that PAA found them both to be higher in the dementia group for subbands higher than 2 Hz is consistent with the fact that SA found higher power in that group for the same subbands (compare Fig. 1 with Figs. 4 and 5). Aside from the possibility that visually well-defined delta waves of frequency lower than 1.6 Hz may indeed be less numerous in dementia, as found in this work, the
FIG. 2. Spectral analysis: comparison of absolute delta power (total and in subbands) in different sleep cycles in control subjects. *P , 0.01 SWS1 versus SWS2, P , 0.01 SWS1 versus SWS3, §P , 0.01 SWS2 versus SWS3. 52
FIG. 3. Spectral analysis: comparison of absolute delta power (total and in subbands) in different sleep cycles in dementia patients. discrepancy between the 2 methods for subbands lower than 1.6 Hz could be because of the inherent differences in the 2 methods, the most important concerning the definition of frequency (Ktonas and Gosalia, 1981; Uchida et al., 1999). In SA, the frequency values relate to the frequencies of the sinusoidal functions, which are used to perform a “best fit” to the morphology of the data window to be analyzed. In PAA, the frequency values relate to the periods (i.e., durations) of the individual waves to be analyzed. As a result, the a priori definition of a wave is necessary in PAA, while it is not needed in SA. Given the above difference in frequency definition, SA can quantify the presence of slow frequencies in the raw data, while these frequencies may not be represented in the raw data as visually well-defined individual waves exhibiting unambiguously observable beginning and ending points in terms of two consecutive local voltage minima, as required for PAA in this work. It could be that in the EEG of SWS in dementia patients, there exists an increase in “diffuse” activity of frequency lower than 1.6 Hz, not exhibiting the form of well-defined delta waves as is the case for delta frequencies higher than 2 Hz, which contributes to the increase in power detected by SA for frequencies lower than 1.6 Hz (compare Fig. 1 with Figs. 4 and 5). Because both methods present advantages and limitations, it has been suggested to apply both SA and PAA to the same set of data (Uchida et al., 1999). Given that power is a function of both wave incidence and wave amplitude, two sets of raw EEG data, which exhibit differences in terms of incidence and amplitude patterns of waves in a given frequency, may lead to the same power estimate at that frequency via SA. In other words, SA may not give any specific information about electrographic detail of the EEG data
FIG. 4. Period-amplitude analysis for the sum of all the SWS periods: mean percent incidence of delta waves in the different subbands in control subjects and patients. *P , 0.0001. Copyright Ó 2012 by the American Clinical Neurophysiology Society
Journal of Clinical Neurophysiology Volume 29, Number 1, February 2012
EEG Delta Characteristics in SWS in Dementia
Issues Related to Pathology In the dementia patients, the delta frequency power did not show the dynamic reduction across NREM/REM sleep cycles (see Fig. 3). This reduction implies a homeostatic mechanism (Darchia et al., 2007) and was found in the control subjects of the present study. However, the reduction between SWS2 and SWS3 in the control subjects was not statistically significant. A comment is necessary on this point because the issue is debated. According to some authors (Darchia et al., 2007; Feinberg and Campbell, 2003), in the elderly, the progressive decline of NREM sleep delta power becomes linear (and not concave-upward curvilinear) and is present across several NREM/REM sleep cycles. Another work (Gaudreau et al., 2001) has shown different changes in delta power across NREM/ REM sleep cycles in middle-aged adults, with a slower slope that seems to approach a plateau value. Our results could be related, at least in part, to the small sample of our population and to the different statistical analysis. It is of interest to observe that the dementia group findings in our work resemble the results from an experimental model for Alzheimer disease shown in transgenic TG 2576 mice (Wisor et al., 2005). According to that study, in TG 2576 mice, compared with wild-type mice, both an increase in sleep EEG power and a disruption of the daily pattern of delta EEG activity were seen, with the disappearance of the progressive decline of the NREM sleep delta power during the light period (i.e., during sleep). The increase in the total delta frequency power during SWS and the modification in the dynamic reduction of delta power across NREM/REM sleep cycles appear to be pathologic manifestations. The observed reduction of well-defined slow delta waves of frequency lower than 1.6 Hz during SWS may be a pathologic sign as well. The role of SWS delta EEG waves in memory processes has been emphasized in recent articles, which have suggested that the level of slow-wave activity might be directly proportional to the strength of cortical synapses related to cognitive processes (Riedner et al., 2007). In particular, the work of Riedner et al. (2007) implies that a reduction in large-amplitude delta waves of frequency lower than 2 Hz during SWS may relate to a decrease in cortical synaptic strength. Assuming that a decrease in cortical synaptic strength may be implicated in the cognitive changes in dementia, the above observation concerning lower than 2-Hz delta waves may help to explain our finding of a reduction in delta waves lower than 1.6 Hz and makes for an interesting hypothesis in dementia, which needs to be investigated further. However, the accentuation of well-defined delta waves of frequencies higher than 2 Hz could also indicate a pathologic finding, and it may relate to compromised brain physiology during sleep in dementia. As a matter of fact, such delta waves may not be
FIG. 5. Period-amplitude analysis for the sum of all the SWS periods: mean amplitude of delta waves in the different subbands in control subjects and patients. *P , 0.0001. it analyzes. However, because PAA can provide specific estimates for wave incidence and wave amplitude, it could quantify the differences in the above two sets of raw data by “decomposing” the power at a given frequency into its two components, incidence and amplitude (Ktonas, 1996; Ktonas and Gosalia, 1981). It can be hypothesized that the amplitude of the EEG waves depends on the mass of cortical neuronal populations generating the EEG via a summation of excitatory and inhibitory postsynaptic potentials, while the incidence of the EEG waves depends on a “pacing” or a “timing” process affecting these populations (e.g., pacing from thalamic centers). Both EEG wave incidence and amplitude could be separately important for EEG analysis, especially in neurodegenerative disorders where cortical degeneration and loss of cortical cells may affect especially the EEG amplitude. Thus, because amplitude and incidence of EEG waves may be differentially affected in neurodegenerative abnormalities, PAA could be particularly useful in searching for EEG-related “biomarkers” in these diseases (Reynolds and Brunner, 1995). Given the preliminary results presented in this work, the use of PAA as an EEG analysis method in dementia appears promising. While PAA has been used to study, among others, the effects of age (Feinberg and Campbell, 2003), homeostatic manipulation (Feinberg et al., 1987), stimulation protocols (Huber et al., 2004), or psychiatric diseases (Armitage et al., 1995) on the amplitude and incidence components of EEG waves, to the best of our knowledge, the importance of PAA has not been assessed in neurodegenerative diseases. More studies, using a larger number of patients representing various kinds of dementia etiologies, and analyzing longer time samples than the ones in this study, need to take place to consolidate the preliminary results presented here.
TABLE 2. Period-Amplitude Analysis: Mean Percent Incidence of Delta Waves in the Different Subbands in Control Subjects (Mean and SD) Versus Related Data of a Single Patient Controls mean 6 SD Controls mean 1 2 SD Dementia (case 1) Dementia (case 2) Dementia (case 3) Dementia (case 4) Dementia (case 5)
0.4–0.8 Hz
0.8–1.2 Hz
1.2–1.6 Hz
1.6–2.0 Hz
2.0–2.4 Hz
2.4–2.8 Hz
2.8–3.2 Hz
3.2–3.6 Hz
31.4 6 30.0 91.4 0.0 0.0 0.1 0.0 0.0
33.8 6 26.2 86.2 0.0 0.0 0.4 0.0 0.8
14.2 6 18.5 51.2 2.0 1.2 2.4 1.7 1.1
8.9 6 13.3 35.5 11.0 6.6 15.3 9.3 8.8
4.9 6 9.5 23.9 13.7 13.8 17.2 15.7 13.8
2.9 6 6.8 16.5 23.9* 23.8* 19.8* 20.8* 19.1*
2.6 6 6.8 16.2 28.6* 30.3* 21.9* 27.3* 27.2*
1.0 6 3.3 7.6 20.7* 23.2* 19.6* 25.1* 28.9*
*Values higher than the value of (mean 1 2 SD) of control subjects.
Copyright Ó 2012 by the American Clinical Neurophysiology Society
53
Journal of Clinical Neurophysiology Volume 29, Number 1, February 2012
E. Bonanni et al.
TABLE 3. Period-Amplitude Analysis: Mean Amplitude of Delta Waves in Different Subbands in Control Subjects (Mean and SD) Versus Related Data of a Single Patient Controls mean 6 SD Controls mean 1 2 SD Dementia (case 1) Dementia (case 2) Dementia (case 3) Dementia (case 4) Dementia (case 5)
0.4–0.8 Hz
0.8–1.2 Hz
1.2–1.6 Hz
1.6–2.0 Hz
2.0–2.4 Hz
2.4–2.8 Hz
2.8–3.2 Hz
3.2–3.6 Hz
100.7 6 119.7 340.1 0.0 0.0 2.1 0.0 0.0
102.5 6 78.3 259.1 0.0 0.0 2.5 0.0 1.7
50.8 6 69.7 190.2 12.2 9.6 10.6 11.0 5.7
30.7 6 50.8 132.3 37.6 26.1 49.0 37.2 17.9
15.8 6 34.0 83.8 40.4 29.1 38.2 39.6 27.6
5.3 6 12.6 30.5 43.1* 27.8 42.1* 45.1* 23.1
4.5 6 11.4 27.3 40.3* 26.3 36.7* 44.4* 25.1
2.5 6 7.6 17.7 32.2* 25.6* 37.3* 42.8* 27.2*
*Values higher than the value of (mean 1 2 SD) of control subjects.
considered as normal sleep delta waves, which are defined as having a frequency of lower than 2.0 Hz (Rechtschaffen and Kales, 1968). Delta waves of higher than 2 Hz could be considered as generated by the same mechanisms responsible for theta EEG activity, which is seen in the EEG of dementia patients at awake (Babiloni et al., 2004; Jeong, 2004) and thus may be viewed as pathologic in this case. Moreover, sleep EEG activity of less than 1 Hz (the so-called slow-wave oscillation) has been associated with the absence of homeostatic regulation and with intracortical neuronal circuits, which may play a major role in organizing delta and sleep spindle activity (Achermann and Borbely, 1997; Steriade et al., 2001) as well as in determining sleep reactivity (Terzano et al., 2005). Because sleep spindle incidence and sleep reactivity are altered in dementia (Crowley et al., 2005; Montplaisir et al., 1995), our findings about a reduction in the incidence and amplitude of well-defined delta waves of frequency lower than 1.6 Hz in dementia seem to be justified.
CONCLUSION In conclusion, computer-based EEG processing via SA and PAA suggests that SWS may be quantitatively altered in dementia as far as its electrographic structure and regulation is concerned. These changes could imply important consequences, which should not be underestimated. Moreover, it should be underlined that the use of computer-based methods could help distinguish normal delta EEG activity from pathologic one in quite a simple way and without stimulation protocols (e.g., Crowley et al., 2005). However, as mentioned already, more studies are needed to validate the preliminary results presented here. REFERENCES Achermann P, Borbely AA. Low-frequency (, 1 Hz) oscillations in the human sleep electroencephalogram. Neuroscience 1997;81:213–222. Ancoli-Israel S, Vitiello MV. Sleep in dementia. Am J Geriatr Psychiatry 2006;14:91–94. Armitage R, Hoffmann R, Fitch T, et al. A comparison of period amplitude and power spectral analysis of sleep EEG in normal adults and depressed outpatients. Psychiatry Res 1995;56:245–256. Babiloni C, Binetti G, Cassetta E, et al. Mapping distributed sources of cortical rhythms in mild Alzheimer’s disease. A multicentric EEG study. Neuroimage 2004;22:57–67. Bliwise DL. Sleep in normal aging and dementia. Sleep 1993;16:40–81. Bliwise DL. Sleep disorders in Alzheimer’s disease and other dementias. Clin Cornerstone 2004;6(suppl 1A):S16–S28. Crowley K, Sullivan EV, Adalsteinsson E, et al. Differentiating pathologic delta from healthy physiologic delta in patients with Alzheimer disease. Sleep 2005;28:865–870.
54
Darchia N, Campbell IG, Tan X, Feinberg I. Kinetics of NREM delta EEG power density across NREM periods depend on age and on delta-band designation. Sleep 2007;30:71–79. Feinberg I, Floyd TC, March JD. Effects of sleep loss on delta (0.3-3 Hz) EEG and eye movement density: new observations and hypotheses. Electroencephalogr Clin Neurophysiol 1987;67:217–221. Feinberg I, Campbell IG. Kinetics of non-rapid eye movement delta production across sleep and waking in young and elderly normal subjects: theoretical implications. Sleep 2003;26:192–200. Gaudreau H, Carrier J, Montplaisir J. Age-related modifications of NREM sleep EEG: from childhood to middle age. J Sleep Res 2001;10:165–172. Huber R, Ghilardi MF, Massimini M, Tononi G. Local sleep and learning. Nature 2004;430:78–81. Jeong J. EEG dynamics in patients with Alzheimer’s disease. Clin Neurophysiol 2004;115:1490–1505. Ktonas PY, Gosalia AP. Spectral analysis vs. period-amplitude analysis of narrowband EEG activity: a comparison based on the sleep delta-frequency band. Sleep 1981;4:193–206. Ktonas PY. Computer-based recognition of EEG patterns. In: Dasheiff RM, Vincent DJ, eds. Continuous Wave-Form Analysis. 1996, pp 23–35. (Electroenceph Clin Neurophysiol Suppl 45). Moe KE, Vitiello MV, Larsen LH, Prinz PN. Cognitive processes and sleep disturbances: Sleep/wake patterns in Alzheimer’s disease: relationships with cognition and function. J Sleep Res 1995;4:15–20. Montplaisir J, Petit D, Lorrain D, et al. Sleep in Alzheimer’s disease: further considerations on the role of brainstem and forebrain cholinergic populations in sleep-wake mechanisms. Sleep 1995;18:145–148. Ohayon MM, Carskadon MA, Guilleminault C, Vitiello MV. Meta-analysis of quantitative sleep parameters from childhood to old age in healthy individuals: developing normative sleep values across the human lifespan. Sleep 2004;27:1255–1273. Rechtschaffen A, Kales A, ed. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Los Angeles, CA: Brain Information Service/Brain Institute, University of California1968. Reynolds CF III, Brunner D. Sleep microarchitecture in depression: commentary. Biol Psychiatry 1995;37:71. Riedner BA, Vyazovskiy VV, Huber R, et al. Sleep homeostasis and cortical synchronization: III. A high-density EEG study of sleep slow waves in humans. Sleep 2007;30:1643–1657. Steriade M, Timofeev I, Grenier F. Natural waking and sleep states: a view from inside neocortical neurons. J Neurophysiol 2001;85:1969–1985. Terzano MG, Parrino L, Smerieri A, et al. CAP and arousals are involved in the homeostatic and ultradian sleep processes. J Sleep Res 2005;14:359–368. Tononi G, Cirelli C. Sleep function and synaptic homeostasis. Sleep Med Rev 2006;10:49–62. Uchida S, Feinberg I, March JD, Atsumi Y, Maloney T. A comparison of period amplitude analysis and FFT power spectral analysis of all-night human sleep EEG. Physiol Behav 1999;67:121–131. Vitiello MV, Bliwise DL, Prinz PN. Sleep in Alzheimer’s disease and the sundown syndrome. Neurology 1992;42((7 suppl 6 42):83–93. Wisor JP, Edgar DM, Yesavage J, et al. Sleep and circadian abnormalities in a transgenic mouse model of Alzheimer’s disease: a role for cholinergic transmission. Neuroscience 2005;131:375–385. Wu YH, Swaab DF. Disturbance and strategies for reactivation of the circadian rhythm system in aging and Alzheimer’s disease. Sleep Med 2007;8: 623–636.
Copyright Ó 2012 by the American Clinical Neurophysiology Society
