Research Article Age-Dependent Increase of Absence ...downloads.hindawi.com/journals/neuroscience/2014/370764.pdf · Research Article Age-Dependent Increase of Absence Seizures and

Research ArticleAge-Dependent Increase of Absence Seizures and IntrinsicFrequency Dynamics of Sleep Spindles in Rats

Evgenia Sitnikova1 Alexander E Hramov23 Vadim Grubov23 and Alexey A Koronovsky24

1 Institute of the Higher Nervous Activity and Neurophysiology of Russian Academy of Sciences Butlerova Street 5AMoscow 117485 Russia

2 Research and Educational Center ldquoNonlinear Dynamics of Complex Systemsrdquo Saratov State Technical UniversitySaratov Polytechnicheskaya Street 77 Saratov 410054 Russia

3 Faculty of Nonlinear Processes Saratov State University Saratov Astrakhanskaya Street 83 Saratov 410012 Russia4 Saratov State University Astrakhanskaya Street 83 Saratov 410012 Russia

Correspondence should be addressed to Evgenia Sitnikova jenia-smailru

Received 28 April 2014 Revised 1 June 2014 Accepted 2 June 2014 Published 23 June 2014

Academic Editor Pasquale Striano

Copyright copy 2014 Evgenia Sitnikova et al This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited

The risk of neurological diseases increases with age In WAGRij rat model of absence epilepsy the incidence of epileptic spike-wave discharges is known to be elevated with age Considering close relationship between epileptic spike-wave discharges andphysiologic sleep spindles it was assumed that age-dependent increase of epileptic activitymay affect time-frequency characteristicsof sleep spindles In order to examine this hypothesis electroencephalograms (EEG) were recorded in WAGRij rats successivelyat the ages 5 7 and 9 months Spike-wave discharges and sleep spindles were detected in frontal EEG channel Sleep spindleswere identified automatically using wavelet-based algorithm Instantaneous (localized in time) frequency of sleep spindles wasdetermined using continuous wavelet transform of EEG signal and intraspindle frequency dynamics were further examined Itwas found that in 5-months-old rats epileptic activity has not fully developed (preclinical stage) and sleep spindles demonstratedan increase of instantaneous frequency from beginning to the end At the age of 7 and 9 months when animals developed maturedand longer epileptic discharges (symptomatic stage) their sleep spindles did not display changes of intrinsic frequencyThe presentdata suggest that age-dependent increase of epileptic activity inWAGRij rats affects intrinsic dynamics of sleep spindle frequency

1 Introduction

Sleep spindles are well-known EEG phenomena that reflectspontaneous rhythmic activity of thalamocortical neuronalnetwork during non-REM sleep [1ndash3] In vivo experimentsdemonstrated a close relationship between sleep spindlesand epileptic spike-wave discharges (SWD) [4ndash6] SWDare electroencephalographic (EEG) manifestation of absenceepilepsy and they are triggered by the cortex oppositeto sleep spindles which are known to originate from thethalamus (reviewed in [7]) In comparison to sleep spindlesSWD are underlain by more intensive excitation andorsynchronization processes in thalamocortical network [7ndash9]

Previously we demonstrated that sleep spindles and SWDshowed different time-frequency characteristics asmeasuredin the cortex and thalamus [10]

Intraspindle frequency is an important parameter char-acterizing intrinsic properties of thalamocortical networkactivity [11 12] with respect to generation of autonomousoscillations In healthy human subjects the frequency ofsleep spindles is known to vary from 10 to 16Hz (eg[2 12 13]) and in animals (rats and cats) from 7 to 14Hz[1 14] Recently we compared time-frequency characteristicsof the anterior sleep spindles in nonepileptic Wistar andepileptic WAGRij rats (genetic model of absence epilepsy)[15] and demonstrated that instantaneous frequency of sleep

Hindawi Publishing CorporationNeuroscience JournalVolume 2014 Article ID 370764 6 pageshttpdxdoiorg1011552014370764

2 Neuroscience Journal

spindles in symptomatic WAGRij rats was constant during aspindle event opposite to ascending dynamics of intraspindlefrequency in control Wistar rats We also found [16] thatsim50 of anterior sleep spindles in WAGRij rats (at the agebetween 5 and 9months) appeared in EEGwith the frequencybetween 8 and 10Hz (mean sim93Hz) 20ndash25 of spindlesmdashwith frequency between 10 and 12Hz (sim114Hz) and 25ndash30mdashbetween 12 and 14Hz (sim134Hz) whereas an increaseof intrinsic frequency during sleep spindle was found at theyounger age (5 months) and only in sim93 and sim114Hz sleepspindles but not in sim134Hz spindles It is well known thatthe number and duration of SWD in WAGRij rats increasewith age [17ndash19] although age-dependent changes in sleepspindles are still uncertain Considering close relationshipbetween epileptic spike-wave discharges and physiologicsleep spindles it was assumed that age-dependent increaseof epileptic activity may affect time-frequency characteristicsof sleep spindles In order to examine this hypothesis inthe current study we compared dynamics of intraspindlefrequency in WAGRij rats at the younger age (preclinicalstate) and elder (symptomatic) ages when SWD are fullydeveloped in EEG

2 Materials and Methods

Experiments were conducted in six male WAGRij rats at theInstitute of Higher Nervous Activity and NeurophysiologyRAS andwere approved by the Ethical Committee on AnimalExperimentation of this Institution EEGs were recordedduring three successive sessions at the age of sim5 months(from 48 to 5) sim7 months (68ndash71) and sim9 months (85ndash9) Rats were equipped with metal screw electrodes thatwere implanted epidurally at the right hemisphere over thefrontal cortex (coordinates AP +2mm and L 25mm relativeto bregma) parietal (AP minus2mm L 55mm) and occipitalareas (AP minus5mm L 3mm) under chloral hydrate anesthesia(325mgkg 4 solution in 09 NaCl) Recordings weremade continuously during a period of 24 hours in freelymoving rats EEG signals were band-pass filtered between 05and 200Hz digitized with 400 samplessecondper channeland stored on hard disk Only frontal EEG data were usedfor time-frequency analysis (because sleep spindles showedmaximum amplitude in the frontal channel) while occipitaland parietal EEGs were used to facilitate determining thestate of vigilance in particular slow wave sleep

Sleep spindles and SWDwere investigated in frontal EEG(Figure 1) for the reason that they both displayed amplitudemaximum in this (anterior) area [5 8 14 17 19] SWD weredetected visually as a sequence of repetitive high-voltagenegative spikes and negative waves that lasted longer than1 sec amplitude of SWD exceeded background more thanthree times [5 17 18] The number and duration of SWDwere scored in 6-hour interval during dark phase Sleepspindles were recognized in EEG as 8ndash14Hz waves withcharacteristic waxing-waning morphology and symmetricalwaveform whose amplitude exceeded background level atleast twice

The continuous wavelet transform (CWT) was used fortime-frequency analysis of sleep spindles in EEG as recordedduring dark phase Three to 5 intervals during slow-wavesleep (duration 15ndash30 s) per animal in each age group wereextractedThe CWT119882(119904 120591) was obtained by convolving theEEG signal 119909(119905) with the basis function 120601

0

119882(119904 120591) =

1

radic119904

int

+infin

minusinfin

119909 (119905) 120601

lowast

0(

119905 minus 120591

119904

) 119889119905

ldquo lowast rdquo denotes complex conjugation(1)

where 119904 time scales (that were converted into Fourierfrequencies 119891) and 120591 the time shift

Complex Morlet wavelet 1206010 was used as basis function

120601

0=

1

4radic120587

119890

119895Ω120578119890

minus12057822 (2)

in which parameterΩ = 2120587Sleep spindles were detected automatically using earlier

developed wavelet-based algorithm in frontal EEG in 15ndash30 sintervals during slow-wave sleep [15 16 20] Briefly waveletenergy119908(119905)wasmeasured in the spindle frequency band 119865 isin(8 16)Hz

119908 (119905) = int

119865

1003816

1003816

1003816

1003816

119882 (119891

119904 119905)

1003816

1003816

1003816

1003816

2119889119891

119904 (3)

The value of 119908(119905) was averaged in the time window 119879

⟨119908 (119905)⟩ = int

119905+119879

119905minus119879

119908(119905

1015840) 119889119905

1015840

(4)

The best quality of automatic recognition was achievedwhen 119879 was set to 05 s that roughly corresponded to theaveraged duration of a sleep spindle Finally the thresholdlevel of wavelet power 119908

119888 was empirically defined and sleep

spindles were identified under condition ⟨119908(119905)⟩ gt 119908

119888 In

order to determine the end point of sleep spindles waveletpower of background EEG was averaged in the frequencyband 119865 isin (8 16)Hz over the time period of 10 s 119908

0 The

value of 1199080was compared with the averaged wavelet power

in the same band 119865 isin (8 16)Hz ⟨119908(119905)⟩ The end pointof sleep spindle was assigned when ⟨119908(119905)⟩ lt 025119908

0 The

true positive detections of sleep spindles reached 90ndash95 ofvisually selected sleep spindles

Rapid changes of the dominant frequency during sleepspindles were explored using ldquoskeletonsrdquo of wavelet surfacesthat were constructed based on the previously describedprocedure [15 16] First the instantaneous wavelet energydistribution 119864

119894(119891

119904 119905

0) = |119882(119891

119904 119905

0)|

2 was computed forthe time moment 119905

0 Second the function 119864

119894(119891

119904 119905

0) was

examined for the presence of local maxima 119864max119896 If severalmaximawere detected in119864

119894(119891

119904 119905

0) the highestmaximumwas

selected and its frequency was considered as the dominantfrequency at the given timemoment 119905

0The value of the dom-

inant frequency was used as the initial point in ldquoskeletonrdquoIn order to plot the full ldquoskeletonrdquo of wavelet surface theabovementioned procedure was repeated for the next time

Neuroscience Journal 3

minus500minus250

0250500

Age = 5m Spike-wave discharges(120583

V)

(a)

minus500minus250

0250500

Age = 5m

(120583V

)

Sleep spindles

(b)

minus500minus250

0250500

Age = 7m

(120583V

)

Spike-wave discharges

(c)

Age = 7mminus500minus250

0250500

(120583V

)

Sleep spindles

(d)


0250500

(120583V

)

1 s


(e)


0250500

(120583V

)

Sleep spindles

(f)

Figure 1 Examples of epileptic activity (spike-wave discharges SWD) and sleep spindles as recorded in frontal EEG in WAGRij rat atdifferent ages Note that at the age of 5 months waveform of SWD was immature (waxing-winning envelope unstable amplitude of spikes ina train)

point Skeletons of wavelet surfaces were constructed in 5 stime intervals containing sleep spindles

Nonparametric Friedmanrsquos ANOVA was used for thestatistical analysis of age-dependent changes of measurableparameters in EEG (with 3 levels of ldquoagerdquo within-subjectdesign repeated measures) and Wilcoxon matched pairs testfor the subsequent post hoc analysis

3 Results

Epileptic activity in WAGRij rats increased with age Thenumber of SWD increased in a period between 5 and 9months of age from 3 to 38 discharges as counted in 6-hourinterval (Friedman test 1205942

119873=6 119889119891=2= 65 119875 lt 005) as well

as the total duration of seizure activity as summed in 6 hours(from 34 plusmn 20 s to 439 plusmn 281 s 1205942

119873=6 119889119891=2= 80 119875 lt 005) At

the age of 5 months SWD were completely absent in 4 out of6 rats and the rest two animals showed very few dischargeswith immature waveform (waxing-winning envelope lowfrequency and unstable amplitude of spikes Figure 1) Thisimplies first that spike-wave activity inMoscowrsquos populationof WAGRij rats appeared at the older age as compared torelatively early development of SWD in native population inNijmegen in which SWD were known to be fully developedat the age of 5 months [18] Second five months of age inour rats might be considered as a preclinical state of absenceepilepsy

In total 115 sleep spindles were automatically selectedin EEG and analyzed in 5-month-old WAGRij rats 117spindlesmdashin 7-month-old and 115 spindlesmdashin 9-month-old Intrinsic frequency dynamics of sleep spindles wereexamined using ldquoskeletonsrdquo of wavelet surfaces constructedin the spindle frequency band 9ndash14Hz It was found thatinstantaneous frequency of sleep spindles slightly changedfrom the beginning to the end of each sleep spindle event

Figure 2 demonstrates three examples of sleep spindles asrecorded in one individual at different ages and correspond-ing wavelet ldquoskeletonsrdquo An increase of dominant frequencywas observed from beginning to the end of a sleep spindleonly at the age of 5months (ascending frequency dynamics inthe ldquoskeletonrdquo plot in Figure 2(a)) but it was no longer presentat the elder ages (Figures 2(b) and 2(c))

For the statistical analysis intraspindle frequency wasexamined in wavelet ldquoskeletonsrdquo by measuring the instanta-neous frequency at the beginning (119891start) and at the end (119891end)of each sleep spindle

It was found first that the value of 119891start increased withage (1205942

119873=115 119889119891=2= 126 119875 lt 0005) Second the value of119891start

at the age of 5 months was lower than that at the age of 7 and9 months (pairwise Wilcoxon test all 119875rsquos lt 005 Figure 3)Third the difference between 119891start and 119891end significantlychanged with age (1205942

119873=115 119889119891=2= 113 119875 lt 0005) According

to pairwise Wilcoxon test ascending intraspindle frequencydynamics that is 119891start lt 119891end were significant only in 5-month-old WAGRij rats (119875 lt 0005) and at the age of 7 and9 months the difference between 119891start and 119891end disappeared(119891start = 119891end Figure 3)

4 Discussion

The current paper demonstrates that intrinsic frequency ofsleep spindles in WAGRij rat model of absence epilepsychanged in parallel to the age-dependent increase of epilepticspike-wave discharges in EEG Five-month-oldWAGRij ratsdeveloped very few SWDwith immaturewaveform thereforethis age has been considered as preclinical stage At thisage intrinsic frequency of sleep spindles increased fromthe beginning to the end Similar elevation of intraspindlefrequency was found previously in nonepilepticWistar rats atthe age of 7 and 9months [15]Therefore in the present studyintraspindle frequency dynamics in 5-month-old WAGRij


9

10

11

12

f(H

z)

05 1 150

Time (s)

100mV

Age 5 months

(a)

f(H

z)

05 1 150

Time (s)

100mV

Age 7 months

9

10

11

12

(b)

9

10

11

12

f(H

z)

05 1 150

Time (s)

100mV

Age 9 months

(c)

Figure 2 Two-second EEG epochs with sleep spindles as recorded in WAGRij rat at three different ages Sleep spindles (marked in grey)were identified automatically based on the continuous wavelet transform Bottom plots demonstrate ldquoskeletonsrdquo of wavelet surfaces in whichdominant frequency fluctuates in the spindle frequency band 9ndash14Hz

ratswere similar towhatwas previously found in nonepilepticWistar [15]

Here it was found that the beginning value of spindle fre-quency (119891start) in preclinical (5months old)WAGRij rats wassignificantly lower as compared to that in older ages (7 and 9months) when epileptic discharges became more numerousand epileptic activity became longer Age-dependent increaseof absence seizures in WAGRij rats is well known fromthe literature [17ndash19 21] Considering the present findingswe can add that age-dependent increase of absence seizureswas associated with changes of intraspindle frequency andthis might be caused by aggravation of epileptic activityin thalamocortical neuronal network due to more intensiveexcitation (hyperexcitation) andor stronger synchronization(hypersynchronization) [7ndash9] Furthermore the low startvalue of the intraspindle frequency 119891start in preclinical (5months old) WAGRij rats might reflect ldquonormalrdquo rhythmicactivity of thalamocortical network

The current results might shed some light on thecontroversial data about interrelation between sleep spin-dles and SWD [7 8] According to the present findingstime-frequency profile of sleep spindles in presymptomaticWAGRij rats was rather normal (and it was similar to age-matched Wistar rats [15]) but it changed in older ageswhen thalamocortical network started producing epilepticdischarges It can be concluded that qualitative changes ofsleep spindles were associated with quantitative changes inSWD It is not surprising because it is well known thatSWD and sleep spindles share the same thalamocorticalpathways (reviewed in [1 7]) Our previous study [21] and

the literature [22] indicated that time-frequency propertiesof SWD changed with age more specifically the frequency ofimmature SWD in younger rats was 5-6Hz and it increasedto 9-10Hz in older animals Taken together both kinds ofthalamocortical oscillations that is sleep spindles and SWDdisplayed age-related changes of time-frequency properties

In order to explain an increase of intraspindle frequencyduring sleep spindles in presymptomatic animals we suggestthe following mechanism It is well known that sleep spin-dles result from mutual interactions between glutamatergicthalamocortical (TC) neurons in specific thalamic nuclei andGABA-ergic cells in the reticular thalamic nucleus RTN(refs in [3]) Neurons in RTN have a propensity to triggerspindle oscillations and fire in bursts at every cycle actingas pacemaker cells TC cells receive inhibitory synaptic inputfrom the RTN and produce rhythmic bursts only once intwo to four cycles Neuronal network mechanism of sleepspindles includes four processes [1 23] (1) initial periodmdashspindle sequence is initiated by the pacemaker cells in theRTN (2) beginning of a spindlemdashsome TC neurons are silentduring the first two-to-four bursts of RTNneurons anddonotreturn signals to the RTN (3) the middle part of a spindlemdashall TC cells burst synchronously with RTN neurons (4)termination of a spindle which is putatively triggered by thecorticothalamic neurons Here in presymptomatic animalswe found that the frequency at beginning of a spindle (thestage 2 of the abovementioned process) was lower than at themiddle and at the end (the stage 3) It seems likely that at thestage 3 more TC cells are recruited by RTN and this resultsin strengthening of rebound inhibition Therefore duration


120

116

112

108

104

100

Inst

anta

neou

s fre

quen

cy o

f spi

ndle

s (H

z)

Age (months)5

7

9

Beginningof spindle (fstart )

Endof spindle (fend )

lowast

Figure 3 Age-dependent changes of the instantaneous frequencyof sleep spindles in WAGRij rats as measured at the beginningof a spindle 119891start and at the end 119891end Asterisk indicates thatinstantaneous frequency in 5-month-old WAGRij rats was lowerthan in older ages and the effect 119891start lt 119891end was significant only in5-month-old WAGRij rats (pairwise Wilcoxon tests all 119875rsquos lt 005)

of each oscillatory cycle becomes shorter and the frequencyof spindle oscillations increased at the end In symptomaticanimals the frequency of spindles at beginning (the stage2) was higher than in presymptomatic animals thereforeall TC neurons in symptomatic rats might be recruited bythe RTN already at the beginning of sleep spindles (similarto the stage 3) and this might lead to the ldquoflatteningrdquo ofintraspindle frequency This putative mechanism ought tobe investigated in the future In general our data indicatethat age-dependent development of absence seizures wasassociated with ldquoflatteningrdquo of intrinsic frequency of sleepspindles The present findings may be helpful for a betterunderstanding of the pathophysiology of absence epilepsyand its probable correlation with sleep disorders particularlywith those related to age

Specific dynamics of intraspindle frequency were alsodescribed in human sleep EEG Changes of intraspindlefrequency (spectral ldquochirprdquo) in humanswere investigatedwiththe aid of Matching Pursuit algorithm [23 24] In particularSchonwald et al [24] studied sleep spindles in C3-A2 EEGchannel in healthy subjects and showed the higher proportionof negatively chirping spindles suggesting that sleep spindlesin healthy humans tend to decelerate their frequency beforetermination In subjects with moderate obstructive sleepapnea there was a decrease in negatively chirping sleep orloss of sleep spindle deceleration This effect was found onlyin slow spindles and only in frontal and parietal regions[25] In the present study flattening of intraspindle frequencydynamics from 119891start lt 119891end in preclinical WAGRij rats to119891start = 119891end in symptomatic animals might be considered

as negative sign that prerequisites development of epilepticspike-wave discharges

We believe that our findings would benefit developmentof new methods for the early (preclinical) diagnosis ofepileptic diseases based on time-frequency properties ofEEG Developing of new strategies acting at preclinical stage(preventing epilepsy in a high-risk group) may be consideredas future directions in research

5 Conclusions

Continuous wavelet transform was used for time-frequencyEEG analysis of sleep spindles in rats with genetic predis-position to absence epilepsy (WAGRij) It was found thatyounger subjects at preclinical stage (5 months old) displayedelevation of intraspindle frequency from the beginning tothe end of sleep spindles but older subjects with fully blownseizures (at the age of 7 and 9 months symptomatic stage)did not display any changes of intraspindle frequency Thisassumes that age-dependent elevation of epileptic activityin WAGRij rats affects intrinsic dynamics of sleep spindlefrequency

Abbreviations

CWT Continuous wavelet transformSWD Spike-wave dischargesTC Thalamocortical neuronsRTN The reticular thalamic nucleus

Conflict of Interests

The authors confirm that there is no conflict of interests ofany sort for any of the authors

Acknowledgments

This study was financially supported by Russian Foundationfor Basic Research (RFBR projects 14-02-31235 13-04-00084and 12-02-31544) and by the Ministry of Education andScience of the Russian Federation (projects SGTU-141 andSGTU-157) The authors thank Professor Ivan N Pigarev forproviding electronic equipment and Dr Elizaveta Rutskovafor technical assistance

References

[1] M Steriade Neuronal Substrates of Sleep and Epilepsy Cam-bridge University Press Cambridge Mass USA 2003

[2] L de Gennaro and M Ferrara ldquoSleep spindles an overviewrdquoSleep Medicine Reviews vol 7 no 5 pp 423ndash440 2003

[3] A Destexhe and T J Sejnowski Thalamocortical AssembliesOxford University Press Oxford UK 2001

[4] P Gloor ldquoGeneralized epilepsy with bilateral synchronous spikeand wave discharge New findings concerning its physiologicalmechanismsrdquo Electroencephalography and Clinical Neurophysi-ology no 34 pp S245ndashS249 1978


[5] E L J M van Luijtelaar ldquoSpike-wave discharges and sleepspindles in ratsrdquoActa Neurobiologiae Experimentalis vol 57 no2 pp 113ndash121 1997

[6] G K Kostopoulos ldquoSpike-and-wave discharges of absenceseizures as a transformation of sleep spindles the continuingdevelopment of a hypothesisrdquo Clinical Neurophysiology vol 111supplement 2 pp S27ndashS38 2000

[7] N Leresche R C Lambert A C Errington and V CrunellildquoFrom sleep spindles of natural sleep to spike and wavedischarges of typical absence seizures is the hypothesis stillvalidrdquo Pflugers Archiv European Journal of Physiology vol 463no 1 pp 201ndash212 2012

[8] E Sitnikova ldquoThalamo-cortical mechanisms of sleep spindlesand spike-wave discharges in rat model of absence epilepsy (areview)rdquo Epilepsy Research vol 89 no 1 pp 17ndash26 2010

[9] A Luttjohann and G van Luijtelaar ldquoThe dynamics of cortico-thalamo-cortical interactions at the transition from pre-ictal toictal LFPs in absence epilepsyrdquo Neurobiology of Disease vol 47no 1 pp 49ndash60 2012

[10] E Sitnikova A E Hramov A A Koronovsky and G van Lui-jtelaar ldquoSleep spindles and spike-wave discharges in EEG theirgeneric features similarities and distinctions disclosed withFourier transform and continuous wavelet analysisrdquo Journal ofNeuroscience Methods vol 180 no 2 pp 304ndash316 2009

[11] T Andrillon Y Nir R J Staba et al ldquoSleep spindles in humansinsights from intracranial EEG and unit recordingsrdquo Journal ofNeuroscience vol 31 no 49 pp 17821ndash17834 2011

[12] Y Urakami A A Ioannides and G K Kostopoulos ldquoSleepspindlesmdashas a biomarker of brain function and plasticityrdquo inAdvances in Clinical Neurophysiology I M Ajeena Ed chapter4 2012

[13] C Iber S Ancoli-Israel A Chesson and S F QuanTheAASMManual for the Scoring of Sleep and Associated Events RulesTerminology and Technical Specification American Academy ofSleep Medicine Westchester Ill USA 1st edition 2007

[14] G Gandolfo L Glin and C Gottesmann ldquoStudy of sleepspindles in the rat a new improvementrdquo Acta NeurobiologiaeExperimentalis vol 45 no 5-6 pp 151ndash162 1985

[15] E Sitnikova A E Hramov V V Grubov andA A KoronovskyldquoTime-frequency characteristics and dynamics of sleep spindlesin WAGRij rats with absence epilepsyrdquo Brain Research vol1543 pp 290ndash299 2014

[16] E I Sitnikova V V Grubov A E Khramov and A AKoronovskiı ldquo[Age-related changes in time-frequency struc-ture of sleep spindles in EEG in rats with genetic predisposi-tion to absence epilepsy (WagRij)]rdquo Zhurnal Vyssheı NervnoıDeiatelnosti Imeni I P Pavlova vol 62 no 6 pp 733ndash744 2012

[17] A M L Coenen and E L J M van Luijtelaar ldquoGenetic animalmodels for absence epilepsy a review of the WAGRij strain ofratsrdquo Behavior Genetics vol 33 no 6 pp 635ndash655 2003

[18] A M L Coenen and E L J M van Luijtelaar ldquoThe WAGRijrat model for absence epilepsy age and sex factorsrdquo EpilepsyResearch vol 1 no 5 pp 297ndash301 1987

[19] G van Luijtelaar and A Bikbaev ldquoMidfrequency cortico-thalamic oscillations and the sleep cycle genetic time of dayand age effectsrdquo Epilepsy Research vol 73 no 3 pp 259ndash2652007

[20] A Ovchinnikov A Luttjohann A Hramov and G van Lui-jtelaar ldquoAn algorithm for real-time detection of spike-wavedischarges in rodentsrdquo Journal ofNeuroscienceMethods vol 194no 1 pp 172ndash178 2010

[21] E I Sitnikova T N Egorova and V V Raevskiı ldquoReductionof the number of neurons in substantia nigra (Pars compacta)positively correlates with a reduction of seizure activity inWAGRij ratsrdquo Zhurnal Vyssheı Nervnoı Deiatelnosti Imeni I PPavlova vol 62 no 5 pp 619ndash628 2012 (Russian)

[22] N Carcak R G Aker O Ozdemir T Demiralp and F Y OnatldquoThe relationship between age-related development of spike-and-wave discharges and the resistance to amygdaloid kindlingin rats with genetic absence epilepsyrdquo Neurobiology of Diseasevol 32 no 3 pp 355ndash363 2008

[23] M Steriade ldquoThe corticothalamic system in sleeprdquo Frontiers inBioscience vol 8 pp d878ndashd899 2003

[24] S V Schonwald D Z Carvalho G Dellagustin E L de Santa-Helena and G J L Gerhardt ldquoQuantifying chirp in sleepspindlesrdquo Journal of Neuroscience Methods vol 197 no 1 pp158ndash164 2011

[25] D Z Carvalho G J Gerhardt G Dellagustin et al ldquoLossof sleep spindle frequency deceleration in Obstructive SleepApneardquo Clinical Neurophysiology vol 125 pp 306ndash312 2014

Submit your manuscripts athttpwwwhindawicom

Neurology Research International

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Alzheimerrsquos DiseaseHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

ScientificaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


BioMed Research International


Research and TreatmentSchizophrenia

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Neural Plasticity


Parkinsonrsquos Disease


Research and TreatmentAutism

Sleep DisordersHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Neuroscience Journal

Epilepsy Research and TreatmentHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Psychiatry Journal


Computational and Mathematical Methods in Medicine

Depression Research and TreatmentHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Brain ScienceInternational Journal of

StrokeResearch and TreatmentHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Neurodegenerative Diseases


Journal of

Cardiovascular Psychiatry and NeurologyHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


spindles in symptomatic WAGRij rats was constant during aspindle event opposite to ascending dynamics of intraspindlefrequency in control Wistar rats We also found [16] thatsim50 of anterior sleep spindles in WAGRij rats (at the agebetween 5 and 9months) appeared in EEGwith the frequencybetween 8 and 10Hz (mean sim93Hz) 20ndash25 of spindlesmdashwith frequency between 10 and 12Hz (sim114Hz) and 25ndash30mdashbetween 12 and 14Hz (sim134Hz) whereas an increaseof intrinsic frequency during sleep spindle was found at theyounger age (5 months) and only in sim93 and sim114Hz sleepspindles but not in sim134Hz spindles It is well known thatthe number and duration of SWD in WAGRij rats increasewith age [17ndash19] although age-dependent changes in sleepspindles are still uncertain Considering close relationshipbetween epileptic spike-wave discharges and physiologicsleep spindles it was assumed that age-dependent increaseof epileptic activity may affect time-frequency characteristicsof sleep spindles In order to examine this hypothesis inthe current study we compared dynamics of intraspindlefrequency in WAGRij rats at the younger age (preclinicalstate) and elder (symptomatic) ages when SWD are fullydeveloped in EEG

2 Materials and Methods

Experiments were conducted in six male WAGRij rats at theInstitute of Higher Nervous Activity and NeurophysiologyRAS andwere approved by the Ethical Committee on AnimalExperimentation of this Institution EEGs were recordedduring three successive sessions at the age of sim5 months(from 48 to 5) sim7 months (68ndash71) and sim9 months (85ndash9) Rats were equipped with metal screw electrodes thatwere implanted epidurally at the right hemisphere over thefrontal cortex (coordinates AP +2mm and L 25mm relativeto bregma) parietal (AP minus2mm L 55mm) and occipitalareas (AP minus5mm L 3mm) under chloral hydrate anesthesia(325mgkg 4 solution in 09 NaCl) Recordings weremade continuously during a period of 24 hours in freelymoving rats EEG signals were band-pass filtered between 05and 200Hz digitized with 400 samplessecondper channeland stored on hard disk Only frontal EEG data were usedfor time-frequency analysis (because sleep spindles showedmaximum amplitude in the frontal channel) while occipitaland parietal EEGs were used to facilitate determining thestate of vigilance in particular slow wave sleep

Sleep spindles and SWDwere investigated in frontal EEG(Figure 1) for the reason that they both displayed amplitudemaximum in this (anterior) area [5 8 14 17 19] SWD weredetected visually as a sequence of repetitive high-voltagenegative spikes and negative waves that lasted longer than1 sec amplitude of SWD exceeded background more thanthree times [5 17 18] The number and duration of SWDwere scored in 6-hour interval during dark phase Sleepspindles were recognized in EEG as 8ndash14Hz waves withcharacteristic waxing-waning morphology and symmetricalwaveform whose amplitude exceeded background level atleast twice

The continuous wavelet transform (CWT) was used fortime-frequency analysis of sleep spindles in EEG as recordedduring dark phase Three to 5 intervals during slow-wavesleep (duration 15ndash30 s) per animal in each age group wereextractedThe CWT119882(119904 120591) was obtained by convolving theEEG signal 119909(119905) with the basis function 120601

0

119882(119904 120591) =

1

radic119904

int

+infin

minusinfin

119909 (119905) 120601

lowast

0(

119905 minus 120591

119904

) 119889119905

ldquo lowast rdquo denotes complex conjugation(1)

where 119904 time scales (that were converted into Fourierfrequencies 119891) and 120591 the time shift

Complex Morlet wavelet 1206010 was used as basis function

120601

0=

1

4radic120587

119890

119895Ω120578119890

minus12057822 (2)

in which parameterΩ = 2120587Sleep spindles were detected automatically using earlier

developed wavelet-based algorithm in frontal EEG in 15ndash30 sintervals during slow-wave sleep [15 16 20] Briefly waveletenergy119908(119905)wasmeasured in the spindle frequency band 119865 isin(8 16)Hz

119908 (119905) = int

119865

1003816

1003816

1003816

1003816

119882 (119891

119904 119905)

1003816

1003816

1003816

1003816

2119889119891

119904 (3)

The value of 119908(119905) was averaged in the time window 119879

⟨119908 (119905)⟩ = int

119905+119879

119905minus119879

119908(119905

1015840) 119889119905

1015840

(4)

The best quality of automatic recognition was achievedwhen 119879 was set to 05 s that roughly corresponded to theaveraged duration of a sleep spindle Finally the thresholdlevel of wavelet power 119908

119888 was empirically defined and sleep

spindles were identified under condition ⟨119908(119905)⟩ gt 119908

119888 In

order to determine the end point of sleep spindles waveletpower of background EEG was averaged in the frequencyband 119865 isin (8 16)Hz over the time period of 10 s 119908

0 The

value of 1199080was compared with the averaged wavelet power

in the same band 119865 isin (8 16)Hz ⟨119908(119905)⟩ The end pointof sleep spindle was assigned when ⟨119908(119905)⟩ lt 025119908

0 The

true positive detections of sleep spindles reached 90ndash95 ofvisually selected sleep spindles

Rapid changes of the dominant frequency during sleepspindles were explored using ldquoskeletonsrdquo of wavelet surfacesthat were constructed based on the previously describedprocedure [15 16] First the instantaneous wavelet energydistribution 119864

119894(119891

119904 119905

0) = |119882(119891

119904 119905

0)|

2 was computed forthe time moment 119905

0 Second the function 119864

119894(119891

119904 119905

0) was

examined for the presence of local maxima 119864max119896 If severalmaximawere detected in119864

119894(119891

119904 119905

0) the highestmaximumwas

selected and its frequency was considered as the dominantfrequency at the given timemoment 119905

0The value of the dom-

inant frequency was used as the initial point in ldquoskeletonrdquoIn order to plot the full ldquoskeletonrdquo of wavelet surface theabovementioned procedure was repeated for the next time


minus500minus250

0250500


V)

(a)

minus500minus250

0250500

Age = 5m

(120583V

)

Sleep spindles

(b)

minus500minus250

0250500

Age = 7m

(120583V

)


(c)


0250500

(120583V

)

Sleep spindles

(d)


0250500

(120583V

)

1 s


(e)


0250500

(120583V

)

Sleep spindles

(f)




3 Results


119873=6 119889119891=2= 65 119875 lt 005) as well


119873=6 119889119891=2= 80 119875 lt 005) At








119873=115 119889119891=2= 113 119875 lt 0005) According


4 Discussion



9

10

11

12

f(H

z)

05 1 150

Time (s)

100mV

Age 5 months

(a)

f(H

z)

05 1 150

Time (s)

100mV

Age 7 months

9

10

11

12

(b)

9

10

11

12

f(H

z)

05 1 150

Time (s)

100mV

Age 9 months

(c)








120

116

112

108

104

100

Inst

anta

neou

s fre

quen

cy o

f spi

ndle

s (H

z)

Age (months)5

7

9



lowast






5 Conclusions


Abbreviations




Acknowledgments


References







































Neural Plasticity










Psychiatry Journal









Journal of



minus500minus250

0250500


V)

(a)

minus500minus250

0250500

Age = 5m

(120583V

)

Sleep spindles

(b)

minus500minus250

0250500

Age = 7m

(120583V

)


(c)


0250500

(120583V

)

Sleep spindles

(d)


0250500

(120583V

)

1 s


(e)


0250500

(120583V

)

Sleep spindles

(f)




3 Results


119873=6 119889119891=2= 65 119875 lt 005) as well


119873=6 119889119891=2= 80 119875 lt 005) At








119873=115 119889119891=2= 113 119875 lt 0005) According


4 Discussion



9

10

11

12

f(H

z)

05 1 150

Time (s)

100mV

Age 5 months

(a)

f(H

z)

05 1 150

Time (s)

100mV

Age 7 months

9

10

11

12

(b)

9

10

11

12

f(H

z)

05 1 150

Time (s)

100mV

Age 9 months

(c)








120

116

112

108

104

100

Inst

anta

neou

s fre

quen

cy o

f spi

ndle

s (H

z)

Age (months)5

7

9



lowast






5 Conclusions


Abbreviations




Acknowledgments


References







































Neural Plasticity










Psychiatry Journal









Journal of



9

10

11

12

f(H

z)

05 1 150

Time (s)

100mV

Age 5 months

(a)

f(H

z)

05 1 150

Time (s)

100mV

Age 7 months

9

10

11

12

(b)

9

10

11

12

f(H

z)

05 1 150

Time (s)

100mV

Age 9 months

(c)








120

116

112

108

104

100

Inst

anta

neou

s fre

quen

cy o

f spi

ndle

s (H

z)

Age (months)5

7

9



lowast






5 Conclusions


Abbreviations




Acknowledgments


References







































Neural Plasticity










Psychiatry Journal









Journal of



120

116

112

108

104

100

Inst

anta

neou

s fre

quen

cy o

f spi

ndle

s (H

z)

Age (months)5

7

9



lowast






5 Conclusions


Abbreviations




Acknowledgments


References







































Neural Plasticity










Psychiatry Journal









Journal of




































Neural Plasticity










Psychiatry Journal









Journal of














Neural Plasticity










Psychiatry Journal









Journal of


Documents

Research Article Age-Dependent Increase of Absence ...downloads.hindawi.com/journals/neuroscience/2014/370764.pdf · Research Article Age-Dependent Increase of Absence Seizures and