- Email: [email protected]
Correlation of electroencephalogram, respiration and movement in the Rett syndrome
Correlation of Electroencephalogram., Respiration and Movement in the Rett Syndrome Alison Kerr, MRCP, David Southall, MD, Patricia Amos, MT, Rosemary Cooper, FRCP, Martin Samuels, MRCP, Jane Mitchell, PhD and John Stephenson, FRCP
Day time video records of 14 girls with the Rett syndrome (RS) (6-17, mean 7 years) were analysed to correlate episodic abnormalities in respiration, movement and electroencephalograms (EEG). Records were compared with those of 12 healthy girls (6-18, median 14 years) who hyperventilated voluntarily. Three RS girls (6-7 years) had minimal respiratory dysrhythmia and showed no correlation between EEG respiration and movement. The other 11 RS girls (6-17 years) had severe awake respiratory dysrhythmia; 10 showing hyperventilation (with hypocapnia) which alternated with active expiratory apnoeic pauses and one with the latter only. All had periods of awake regular breathing with normal respiratory gases. In these girls EEG showed non-epileptic generalised slow activity some of which was paroxysmal. In the six youngest (6-10 years) of these 11 RS girls, non-epileptic paroxysms of EEG slow activity at 1%-4 Hz occurred and were associated with periods of normal breathing and normal pC0 2 levels whether girls were alert, drowsy or asleep, but were uncommon during episodes of hyperventilation (and hypocapnia). In four of these girls the EEG paroxysms occupied ~ 1-3% of the time during periods of respiratory dysrhythmia (81 minutes) and 8-100% of the time during alert periods with normal breathing (29 minutes), p ~ O. 001 for this difference. Short bursts of EEG slowing occasionally followed prolonged apnoeic pauses. In two cases brief partial complex seizures occurred. In five of these girls stereo typic movements exacerbated during episodes of respiratory dysrhythmia and reduced during normal breathing. In four clear demarcation of these changes allowed comparison of time spent in vigorous stereotyped movement during periods of awake normal and dysrhythmic breathing (p ~ 0.01 for this difference). In the older girls (11-17 years) stereotyped movements did not fluctuate with periods of respiratory dysrhythmia. Fairly per~stent generalised, largely unreactive theta activity at 4-6 Hz was present in all these girls and in one tended to increase at the end of apnoeic pauses. The consistent appearance in the younger girls of episodes of non-epileptic EEG slow activity during normal breathing, their relative absence during hyperventilation and the association between episodes of respiratory dysrhythmia and exacerbation of stereo typic movement helps to define the underlying neurological mechanisms in the Rett syndrome. Key words: Rett syndrome, respiratory dysrhythmia, hyperventilation, apnoea, electroencephalogram (EEG) Valsalva, stereotypy, episodic abnormality, non-seizure EEG paroxysms. Kerr A, Southall D, Amos P, Cooper R, Samuels M, Mitchell J, Stephenson J. Correlation of electroencephalogram, respiration and movement in the Rett syndrome. Brain Dev 1990;12:61-8
From the Department of Child Health (AK) and Fraser of Allander Unit (JS), Royal Hospital for Sick Children, Yorkhill, Glasgow; Monitoring Unit (AK) and Epilepsy Centre (PA), Quarrier's Homes, Renfrewshire; Department of Paediatrics, Cardiothoracic Institute, London (DS); Department of Postgraduate Medicine, North Staffordshire Hospital Centre, Staffordshire (RC); Department of Paediatrics, National Heart and Lung Institute, Brompton Hospital, London (MS); Department of Mathematics, Strathc1yde University, Glasgow (JM). Correspondence address: Dr. Alison Kerr, Monitoring Unit, Quarrier's Homes, Bridge of Weir, Renfrewshire, PAll 3SA, Scotland UK. The investigation was carried out at the Epilepsy Centre, Quar-
The Rett syndrome (RS) [1, 2] is associated with profound mental and physical disability in 1 in 10-12,000 girls [3,4] who may survive into the fourth decade. Signs appear in the first year of life [3-6] and exacerbate at 1-2 years. There are deficits in mental processing, extrapyramidal [3], pyramidal [7,8] and respiratory function. Diagnosis is easiest in the child between 5 and 15 years. rier's Homes, Bridge of Weir, Renfreshire, PAll 3SA in July 1987 and February 1988. Presented to British Paediatric Neurology Association in January 1988.
This is partly due to the occurrence of well defined episodes which include bursts of hyperventilation alternating with apnoeic pauses, exacerbations of stereotyped movements most marked in the face and limbs, non-seizure vacant spells and abnormal paroxysmal EEG activity.
mixer
ANV\
I-_:--:--:-l
f---+-
exp
timer tape
Resp
t--_ _ _--t'esplt'ace
97
02
o
In order to study these phenomena we constructed the apparatus shown in Fig 1 [9, 10] which records simultaneously the required parameters in the awake and unrestricted child. This is the second of two papers presenting our results. In our first paper [10] we described the main respiratory observations in 18 girls with RS (case numbers for that study are given in brackets in Table 1). Ten girls (1-10 of that paper) hyperventilated only when awake with development of severe hypocapnia. Hyperventilation was interspersed with prolonged apnoeic pauses (> 19 secs) with hypoxia in 47% of pauses. A further four patients who had hyperventilated in the past no longer hyperventilated but had frequent apnoeic pauses (cases 11-14 of that paper). The remaining four cases showed only occasional apnoeic pauses with Valsalva manoeuvres and were thought never to have hyperventilated. The aim of the present study was to establish the relationships between the non-seizure episodic events in EEG breathing and movement in 14 of the same girls and to observe changes with age.
PATIENTS AND METHODS Patients are listed in Table 1 with the essential respiratory
Fig 1 Method for recording behaviour. respiration and EEG.
Table 1 Fourteen Rett syndrome cases: respiratory. EEG and movement data Case number
1*5 (9)
2 (1)
Age (years)
6
6
Hypervenfilation present
+
+
(
*1 )
Lowest CO. recorded *2
3*6 (10) 6 +
4 (5)
5 (7)
6 (4)
7 (3)
8 (2)
9 (8)
10 (6)
7
7
10
11
12
13
16
+
+
+
+
+
+
+
1.8/18 2.2/24 2/13 1.4/13 1.6/8 2.1/18
11 (11 )
17 _*9
4.2/33 2.2/26 2.4/26 2.2/20 3.8/NR
]2*7 13*8 (15)
(18)
14 (16)
6
6
7
4.5/44 4/37 4.4/46
Apnoeic pauses present*3
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Lowest oxygen recorded*4
<50
60
<50
<50
70
<50
92
<50
92
66
80
94
98
97
EEG paroxysms at 1%-4 Hz
+
+
+
+
+
Increased in normal CO 2
+
+
+
+
+
4-6 Hz EEG activity
+ UR
+ UR
+ UR
+ R
+ UR
Paroxysmal 4-6 Hz activity Paroxysmal limb movement with resp. dysrhythmia
+
+
+
+
+
+ UR +*10
+ R
+ UR +
+
+ UR
+ UR
+ UR
+
+
+ UR
+
+
*1) Brackets after case numbers give case numbers for our fIrst paper. *2) Figure above the line is end-tidal carbon dioxide in volumes % (normal> 4); below the line is transcutaneous carbon dioxide mmHg (normal> 35 mm). * 3) Active expiratory apnoeic pauses, all included Valsa1va-like manoeuvers. *4) % saturation (normal> 97%). *5, *6, *7, *8) Cases with minor atypical features; *5) Better than expected use of hands. *6) Threatened miscarriage in pregnancy. *7 and * 8) Earlier stereotypic hand movements had almost disappeared. *9) Hyperventilation recorded in the past. *10) Abnormality increased when end-tidal CO. was normal. UR: unreactive EEG, R: reactive EEG. Anticonvulsant drugs: Carbamazepine (cases 2,3,5,8,9,11,12), sodium valproate (cases 6, 14), c10nazepam (cases 5, 13). NR: not recorded.
62 Brain & Development. Vol 12, No 1, 1990
data from our previous paper and the results of the present study. All girls were diagnosed as having RS and no other disorder. Minor atypical features in four girls are described in the footnote of the Table. Four girls included in our first paper were excluded because EEG was not simultaneously recorded. Twelve healthy girls, sisters and friends of RS patients were also studied (age range 6-18, median 14). Monitoring equipment Respiratory equipment: Full details are given in our previous paper [10] non-invasive respiratory monitors measured transcutaneous pC0 2 (Hewlitt Packard) end-tidal CO 2 (Engstrom) breathing movements by inductance plethysmography (respitrace, Studley data systems) and arterial oxygen saturation (Sa02) by pulse oximeter (Nellcor, modified to provide a beat to beat Sa0 2). The plethysmographic waveforms from the pulse oximeter were monitored to verify Sa02' All end-tidal CO 2 values were required to have adequate end-expiratory plateaus. Electroencephalographic recording: Seven bi-polar channels of EEG and one of ECG were recorded on a battery operated tape recorder using collodion applied scalp electrodes and chest electrodes with pre-amplifiers on a collar round the neck (medilog 9000, Oxford Medical Systems). EEG electrodes were applied in a modified montage in order to provide maximum information, bearing in mind the range of the child's activity as well as the need to correlate with other parameters. The EEG and ECG signals were relayed through a monitor/writer/coupling unit to a conventional EEG machine (Nihon Kohden). In all EEG recordings sensitivity was 7.5 microV per mm, time constant 0.3 secs with high frequency filter at 70 Hz. Electrode impedance was maintained between 2-5 Kohms. Video/audio recording: A portable camera (Sony video 8) was used for all recordings. A video mixer (ForA FVG 600) transmitted the EEG and the video picture of the subject, with seconds time trace, to a monitor screen and video recorder. Respiratory data was scanned directly from the individual respiratory monitors which were stacked beside the subject. Respiratory data was also stored independently on audio tape. Girls were monitored in daytime for periods of one to six hours (median 3). Seven RS girls slept for some part of the recording. In girls with severe hypocapnia due to hyperventilation, the carbon dioxide level was raised towards normal values (end-tidal 4.5 vol %, TcPC0 2 35 mmHg) by rebreathing inside a perspex hood or a mask. This did not cause distress and base line arterial oxygen saturations remained normal (> 97%). The twelve healthy girls, similarly recorded, were asked to hyperventilate in air and on separate occasions in the rebreathing hood.
Data extraction from records Abnormal episodes in respiration, EEG and movement were identified from video records. Such episodes were required to be distinctive and contain the same sequence of abnormal events on each occurrence. Episodes were defined independently for each girl. Episodes in behaviour were sequences of skeletal movement affecting trunk, face and limbs, identical on repeated occasions. Levels of this activity were defined for each girl (++, + or -). In addition to episodes of movement a record was made of interactions involving the girl. Respiratory episodes were defined as follows: 1) Hyperventilation: increase in respiratory excursion with subsequent fall in transcutaneous and end expiratory CO 2 level). 2) Apnoeic pause/s: pause or succession of pauses in respiratory excursion. 3) Valsalva-like manoeuvre: straining vocalisation in the course of apnoeic pause/so 4) Normal breathing: regular chest movement with rising CO 2 levels. "Other breaths" included irregular breaths other than those defined. Respiratory dysrhythmia is the term used to describe periods consisting of alternating episodes of hyperventilation and apnoeic pauses. Blood gas levels were confirmed from audio tape recordings. Electroencephalograms were analysed visually with particular reference to the distribution and responsiveness of rhythmic and/or paroxysmal activity and the presence or absence of paroxysmal slowing and epileptogenic activity. Non epileptic 'paroxysmal activity was characterised by sudden increase in amplitude of slow components (delta and theta) lasting more than 2 secs. The state of each patient was recorded as asleep, awake or transitional. Episodes from each area of function were viewed "blind" by covering the parts of monitor screen and chart showing other areas of function. EEG records and respiratory records were reviewed independently by individual authors. All the data from each patient's video tape was charted on continuous paper which was divided into columns for each area of function in which episodes and states were to be correlated. Time was marked at one horizontal line for 4 seconds. Onset and end of episodes were recorded in seconds. In cases where episodes were sharply defined, the numbers and durations of each type of episode and state and the associations between them were calculated. Other records were reported in the conventional manner with events in other parameters being taken into account in the reporting of each.
RESULTS Patients fall into 3 groups according to results (Table 1).
Kerr et al: EEG respiration and movement in Rett syndrome 63
Group 1 (Six girls, cases 1-6,6-10 years)
All these girls had episodes of severe hyperventilation with hypocapnia and alternating apnoeic pauses and all showed non-epileptic paroxysmal slow activity on EEG. These paroxysmal EEG changes were generally asso· ciated with periods of normal breadling (and normal CO 2 levels) whether the girls were alert, drowsy or in day·time sleep . They seldom occurred during episodes of hyperventilation. In cases 1-4 the paroxysmal EEG change consisted of 1~-4 Hz activity of 75-250 microV in amplitude. In case 5 it consisted of slow spike and wave discharges lasting 10-30 sees and too-I 50 microV in amplitude. In case 6 the change consisted of paroxys· mal bursts of 4-6 Hz activity of 50-75 microV amplitude. Fig 2 illustrates the relatively "normal" EEG seen during HV and the increase in abnormality which occurred during normal awake breathing in cases 4 and 5. Fig 3 shows the parallel abnormalities in EEG and respiration
in case 5. For cases 1, 2, 4 and 5 in whom episodes were most clearly demarcated, seizure-free periods were selected to allow comparison beween the frequency of these episodes in periods of dysrhythmic breathing, normal breathing (alert state) and normal breathing (drowsy or asleep). During periods of respiratory dysrhythmia, hypocapnia was always present. Within these selected periods significant hypoxia < 50% was recorded on one occasion without alteration in EEG or behaviour. Table 2 compares the time occupied by the nonepileptic EEG paroxysms during periods of established normal and dysrhythmic breathing during the selected periods in these fou r girls. Chi-squared tests were carried out in cases I, 2 and 5 to compare regular discrete EEG paroxysms in the various periods of breathing. There was a clear difference in the number of buTSis during respiratory dysrhythmia (Jow) and normal breathing
•
• • ••
,~
TC
'----------------------------------------I~·-----------------------------------
0.> _
~ .~
7O.vL-
v..__t" ...:.,_._.!~~
~jv:Jl..·p.;A.--!;t..~.IWA-..-Jt,.'.I\,.-J~ {~L,.J.".,..1r!.r. .~/'-·~,.
'-
~V~'V'C/_ ' \J,""''-,v\i\A/'vv-.'u\ir;\.-\rL;''.~v7 ~/""f\-NI..rl'·;v.r.Nf'>JV'c.l'-I'
,
-~~~V""""'-~-
~~~ .r'""r'v~;II.,'i\I'r.J" f,!,"__f'r't,('iY/Vy ·~/~MV'__('.rvv..rvf'·Iv-. . ,...,.r-..~/\i\/V. ..r_v
,~
Tt
0.3_
• ••
• ~\.----~~.,.,r..-..
CH HOC!
~~Iv-.~/.-'. A...IvI'A--A--~~A.-..~~.A~vA~V-----~~--~--~~~~~v~
~~----~~~~~~~~--~~~~
•
,~
,,~
l .C. u _
• •• ~,~
•
"""
L
IO.VL -
,
'-
•
I
t'~.
., I
Fig 2 Electroencephalograms in cases of Rett ryndrome. EXQmpJes sho w the EEG in cares 4 ( ) arul j ( oj and c); and dun'ng normal breathing b) and d). Example e) sho ws the theta rhythm in case 10 ( j.
64 Brain
&;
Development, Vol 12. No 1,1 990
)
during hyperventilation
JTTTITTTl""'~'TTTTTTTT'J"TTTTTTTTLJ'TTTTTTTT'~~
~V\')~\I
i~~j~~Jr~~
I
i~~"'~'~lr~
~""j~,.r""~ i
r~
~
. "
f" "
r. ~ f
, I . . . ,. ~ ... } ,~
rr r ~
,
~
, ( f '.• t ... ' !
~
r
~
i1,
.. ' , I
f
Fig 3 Parallel records of EEG breathing and behaviour in case 5. 1, 2 and 3 recordings during a period of respiratory dysrhythmia. 4, 5 and 6 recordings during alert normal breathing (while rebreathing). Small letters indicate respiratory activity corresponding to the photographs. In case 5 the paroxysms of movement ++ occurred not only during periods of respiratory dysrhythmia but specifically during the HVepisodes within these periods. During normal breathing, movements diminished.
(high), p = 0.001 for this difference. The number of bursts occurring in alert and asleep periods of normal breathing was not significantly different. In cases 1, 2, 4 and 5 the EEG paroxysms occupied 1-3% of the time during respiratory dysrhythmic periods and 8-100% during alert normal breathing periods (p~O.OOl for this difference). Episodes of stereotyped movement involved face trunk limbs and hands, and were graded ++ when they were very energetic with wringing or clapping of the hands, + when they were intermediate with tense posturing of the limbs and - when absent or minimal with fingertip movement. In cases 1, 2, 4, 5 and 6 episodes of stereotyped move-
ment (++ and +) occurred during the periods of respiratory dysrhythmia and diminished (+ and -) during periods of normal breathing (see Table 2). The percentages of ++ activity for the nineteen continuous periods recorded were compared using a non parametric test (KruskalWallis) which assumes that the various periods are independent. This test shows a significant difference in amount of ++ movement between dysrhythmic and normal alert breathing periods (p = 0.01). In case 5 the episodes of vigorous movement (++) coin. cided precisely with the episodes of hyperventilation and during apnoeic pauses reduced instantaneously to the less intense level (+), so that limb and chest movements oc-
Kerr et al: EEG respiration and movement in Rett syndrome 65
Table 2 Comparison of non-epileptic EEG paroxysms occurring during periods of normal and dysrhythmic breathing in four cases
Case
Time sec
ETC0 2 vol%
no
EEG bursts %o/time
Respiratory dysrhythmia: 741 (2) 1.9-3 1 1 2.8-3.5 1,784 (4) 14 2 2-3 2 1,108 (3) 4 2-3 1,264 (3) 5 Normal breathing: alert 4-5 124 (2) 5 1 4.9 239 (1) 9 2 3.8-4 389 (2) 4 Contino 3.1-4.4 985 (2) 36 5 Normal breathing: asleep or drowsy 4-5 448 1 25 4-5 623 2 40 4 3.7-5 Contino 597 4.7-4.9 150 5 9
Movement % o/time
++
<1 3 2 <1
92 38 63 54
8 8 100 57
0 0 0 0
+
3 56 35 46
5 6 3 0
48 52 22 78 32 68 98 2
16 24 61 29
In case 5 movement ++ coincided with HV and + with apnoeic pauses. ( ) Brackets indicate the numbers of periods analysed.
curred exactly together. Fig 3 illustrates these events. An agitated manner accompanied stereotyped movement ++ and a preoccupied expression appeared during Valsalva-like manoeuvers. In case 3 the stereotyped activity was not episodic. The change from a period of normal breathing to respriratory dysrhythmia was seen when a girl woke spontaneously, was awakened, was offered an object which captured her interest, or was spoken to directly. At these' times the episodes of respiratory and limb dysrhythmia commenced simultaneously giving the impression of a burst of movement outwith her control but linked to her alerting. The change from periods of respiratory dysrhythmia to normal breathing was recorded when a girl became drowsy, following a period of hiccoughs or crying and when rebreathing had raised the level of inspired CO 2 towards normal. Group 2 (Five girls, cases 7-11, aged 11-17 years) These were the remaining girls with severe respiratory dysrhythmia. In cases 7 to 10 this included hyperventilation with hypocapnia and in case 11 only the apnoeic pauses with Valsalva-like manoeuvers. In these girls the regular stereotyped limb activity was not paroxysmal. Valsalva-like manoeuvers were accompanied by an anxious appearance. On EEG all five girls showed, as well as some polymorphic delta activity, prominent 4-6 Hz theta activity which was often synchronous and generalised. This was largely unreactive to normal physiological stimuli but did
66 Brain & Development, Vol 12, No 1, 1990
show some variation in amount and amplitude from time to time. In cases 8 and 11 the 4-6 Hz activity showed paroxysmal increase in amplitude towards the end of apnoeic pauses. Fig 2 illustrates the persistent rhythmic 4-6 Hz activity in case 10. Group 3 (Three girls, cases 12-14, aged 6-7 years) These girls showed only occasional apnoeic pauses with Valsalva-like manoeuvers and normal gases and had no history of hyperventilation. Their stereotyped limb movements were not episodic. Case 14 had paroxysmal slow activity at 1~-4 Hz (in association with spike discharges) which fluctuated in amount and was particularly prominent during sleep. Seizures and non-seizure vacant spells Six children (cases 1, 2, 3,4, 5 and 6) had attacks of vacancy and staring not associated with significant EEG changes. Cases 1 and 2 each also had a minor complex seizure, with appropriate ictal activity on EEG during attacks. Both the non-seizure spells and the seizures occurred after periods of intense hyperventilation with severe hypocapnia and were associated with apnoeic pauses with or without a Valsalva-like manoeuver. Pulse oximetry indicated subsequent but not prior hypoxaemia. Nine of the 14 RS girls (cases 1, 2, 5, 6, 7, 8, 9, 12 and 14) showed some epileptogenic activity in the form of single or grouped spike and sharp wave discharges or slow spike and wave, distinct from the episodes of paroxysmal slow activity. Our data on seizure and non-seizure vacant spells will be the subject of a separate publication. All the RS girls showed general EEG abnormality with excess of polymorphic slowing and poorly developed (daytime) sleep change. Case 4 showed least abnormality although the EEG was poorly developed. The 12 healthy girls hyperventilated voluntarily for 3.9-15.8 minutes (median 7.8). Carbon dioxide levels dropped to 1.9-3.3 vol% (median 2.1). Nine (all those under 15) showed a normal slow wave response to this hypocapnia. The three who failed to show a normal response were aged 18, 15 and 18 years and developed CO 2 levels of 3.3,2.4 and 1.9 vols% respectively. None of the girls showed limb or face movement comparable to that of the RS girls.
DISCUSSION The striking observations from this study are the association of episodic non-epileptic paroxysmal abnormality on EEG with periods of normal breathing; and the association of the episodes of repetitive movement with the respiratory dysrhythmia. These are so consistent in the younger girls as to provide useful diagnostic signs. As with other features of RS these seem related to a particular age period between about 4 years (following
regression) and the early teens, after which dysrhythmic breathing is less inclined to include hyperventilation with hypocapnia and both the limb movements and EEG changes may become less episodic. The occurrence in these very disabled girls, of such complex stereotyped paroxysms of movement involving face trunk and limbs simultaneous with the respiratory dysrhythmia seems to suggest a common trigger for both which would appear to be outwith the voluntary control of the child although possibly linked to alerting. In cases 1, 2, 4 and 5 it is notable that the episodic non-epileptic paroxysmal EEG activity occurred when breathing and respiratory gases were normal, whether after waking from normal sleep, during daytime sleep, during rebreathing or at other times, while on the other hand, this paroxysmal activity was seldom seen during hypocapnia and none of the RS girls showed a normal slow wave response to hyperventilation. In normal children, particularly between 6 and 12 years slow wave activity develops during hyperventilation, associated with hypocapnia [11]. Absence of this response is not abnormal but would be unusual with significant hypocapnia, especially in a younger age group. This normal slow wave response was seen in 9 of our 12 healthy girls including all the youngest. The mechanisms underlying normal EEG changes during hyperventilation with hypocapnia are poorly understood in spite of a good deal of work having been done over many years. The subject has been carefully reviewed by Patel and Morsby [12] although they note that their conclusions are speCUlative. We cannot equate the paroxysmal abnormality of the RS girls with the normal EEG response to hyperventilation since the amplitudes and wave forms are different. Our findings therefore remain to be explained. The nature of these changes and their presence during drowsy and asleep states as well as in the apparently alert child might indicate that they reflect the existence of a pathological "non-alert" state, a consequence of the essential developmental deviation responsible for the syndrome. Such a state might be expected to provoke an extreme alerting reaction. The prominent 4-6 Hz activity is an interesting feature of the RS cases. This has been described [13-15] but has been considered a feature of slowing with increasing age. On the contrary it was present in our youngest girls, appeared reactive in case 4, the child with the least abnormal record and in case 7, while in the other cases it was unreactive to physiological stimuli and behavioural change although fluctuating from time to time. In four cases [1,6,8,11] it showed occasional paroxysmal increase in amplitude. This theta activity which is a normal feature of childhood appears in RS girls to develop as expected, then to become unresponsive and finally totally unreactive and persistent. It seems to be highly characteristic of these
girls. In our series slow activity in the delta range, whether polymorphic, rhythmic and/or paroxysmal, was more prominent in the younger age groups. In this paper we have not fully presented or discussed data on the important relationships between seizures and the respiratory dysrhythmia. However our observations of partial complex seizures and non-seizure vacant spells indicate that without EEG monitoring it is difficult to distinguish minor seizures from other attacks and that both kinds of attack may relate to the respiratory abnormalities. We now plan to explore further both seizure and nonseizure vacant spells, to investigate the respiratory control mechanisms in RS, awake and asleep and to observe the effects of morphine antagonists on the episodic phenomena which we have documented. It is no surprise that a pathological process producing profoundly disordered thinking and movement may also involve respiratory control. The assumption that RS may be due to a single gene defect makes the neural mechanisms underlying these disordered functions of peculiar interest.
ACKNOWLEDGMENETS The authors gratefully acknowledge the technical assistance of Mr. E. Picton Jones, Dr. A. Etchells, and staff of the medical illustration department RHSC also the support of medical colleagues and the families of subjects. Dr. Kerr is funded by Quarrier's Homes, the Rett Syndrome Associations (National and UK) and Dr. Southall and Dr. Samuels by the National Heart and Chest Hospital. Equipment was funded by the Scottish Society for the Mentally Handicapped, National Rett Syndrome Association, Labaz-Sanofi, Oxford Medical and Nellcor.
REFERENCES 1. Rett A. Uber ein eigenartiges hirnatrophisches Syndrom bei Hyperammonamie im Kindsalter. Wien Med Wochenschr 1966;116:723-6. 2. The Rett Syndrome Diagnostic Criteria Work Group. Diagnostic criteria for Rett syndrome. Ann Neurol 1988;23: 425-8. 3. Kerr A, Stephenson J. A study of the natural history of Rett syndrome in 23 girls. Am J Med Genet 1986;24(suppll): 77-83. 4. Hagberg B. Rett syndrome: Swedish approach to analysis of prevalence and cause. Brain Dev (Tokyo) 1985;7:277-80. 5. Nomura Y, Segawa M, Hasegawa M. Rett syndrome - clinical studies and pathophysiological consideration. Brain Dev (Tokyo) 1984;6:475-86. 6. Kerr A, Montague J, Stephenson J. The hands and the mind pre- and post-regression in Rett syndrome. Brain Dev (Tokyo) 1987;9:487-90. 7. Eyre J, Kerr A, Miller S, O'Sullivan M, Ramesh V. Neurophysiological observations on the corticospinal projections to the upper limb in subjects with Rett syndrome. (submitted for publication) 8. Oldfors A, Hagberg B, Nordgren H, Sourander P, Witt-Enger-
Kerr et al: EEG respiration and movement in Rett syndrome 67
strom 1. Rett syndrome: spinal cord neuropathology. Pediatr Neural 1988;4:172-4. 9. Kerr A, Amos P, Etchells A, Irwin A, Holmes T, Stephenson J. A low cost method for simultaneous video recording of ambulant subject and electroencephalograph: the Quarrier's system. J Ment Defic Res 1988;32:497-500. 10. Southall D, Kerr A, Tirosh E, Amos P, Lang M, Stephenson J. Hyperventilation in the awake state: potentially treatable component of Rett syndrome. Arch Dis Child 1988;63: 1039-48. 11. Binnie C, Coles P, Margerison J. The influence of end-tidal carbon dioxide tension on EEG changes during routine hyperventilation in different age groups. Electraencephalagr Clin Neuraphysial 1969;27:304-6.
68 Brain & Development, Vol 12, No 1, 1990
12. Patel V, Maulsby R. How hyperventilation alters the electroencephalogram: a review of controversial viewpoints emphasizing neurophysiological mechanisms. J Clin NeuraphysiaI1987;2:101-20. 13. Niedermeyer E, Rett A, Renner H, Murphy M, Naidu S. Rett syndrome and the electroencephalogram. Am J Med Genet 1986;24(suppl1):195-9. 14. Verma N, Cheda R, Nigro M, Hart Z. Electroencephalographic findings in Rett syndrome. Electroencephalagr Clin Neuraphysial 1986 ;64: 394-40 1. 15. Hagne I, Witt-Engerstrom I, Hagberg H. EEG development in Rett syndrome. A study of 30 cases. Electraencephalagr Clin NeuraphysiaI1989;72:1-6.
Day time video records of 14 girls with the Rett syndrome (RS) (6-17, mean 7 years) were analysed to correlate episodic abnormalities in respiration, movement and electroencephalograms (EEG). Records were compared with those of 12 healthy girls (6-18, median 14 years) who hyperventilated voluntarily. Three RS girls (6-7 years) had minimal respiratory dysrhythmia and showed no correlation between EEG respiration and movement. The other 11 RS girls (6-17 years) had severe awake respiratory dysrhythmia; 10 showing hyperventilation (with hypocapnia) which alternated with active expiratory apnoeic pauses and one with the latter only. All had periods of awake regular breathing with normal respiratory gases. In these girls EEG showed non-epileptic generalised slow activity some of which was paroxysmal. In the six youngest (6-10 years) of these 11 RS girls, non-epileptic paroxysms of EEG slow activity at 1%-4 Hz occurred and were associated with periods of normal breathing and normal pC0 2 levels whether girls were alert, drowsy or asleep, but were uncommon during episodes of hyperventilation (and hypocapnia). In four of these girls the EEG paroxysms occupied ~ 1-3% of the time during periods of respiratory dysrhythmia (81 minutes) and 8-100% of the time during alert periods with normal breathing (29 minutes), p ~ O. 001 for this difference. Short bursts of EEG slowing occasionally followed prolonged apnoeic pauses. In two cases brief partial complex seizures occurred. In five of these girls stereo typic movements exacerbated during episodes of respiratory dysrhythmia and reduced during normal breathing. In four clear demarcation of these changes allowed comparison of time spent in vigorous stereotyped movement during periods of awake normal and dysrhythmic breathing (p ~ 0.01 for this difference). In the older girls (11-17 years) stereotyped movements did not fluctuate with periods of respiratory dysrhythmia. Fairly per~stent generalised, largely unreactive theta activity at 4-6 Hz was present in all these girls and in one tended to increase at the end of apnoeic pauses. The consistent appearance in the younger girls of episodes of non-epileptic EEG slow activity during normal breathing, their relative absence during hyperventilation and the association between episodes of respiratory dysrhythmia and exacerbation of stereo typic movement helps to define the underlying neurological mechanisms in the Rett syndrome. Key words: Rett syndrome, respiratory dysrhythmia, hyperventilation, apnoea, electroencephalogram (EEG) Valsalva, stereotypy, episodic abnormality, non-seizure EEG paroxysms. Kerr A, Southall D, Amos P, Cooper R, Samuels M, Mitchell J, Stephenson J. Correlation of electroencephalogram, respiration and movement in the Rett syndrome. Brain Dev 1990;12:61-8
From the Department of Child Health (AK) and Fraser of Allander Unit (JS), Royal Hospital for Sick Children, Yorkhill, Glasgow; Monitoring Unit (AK) and Epilepsy Centre (PA), Quarrier's Homes, Renfrewshire; Department of Paediatrics, Cardiothoracic Institute, London (DS); Department of Postgraduate Medicine, North Staffordshire Hospital Centre, Staffordshire (RC); Department of Paediatrics, National Heart and Lung Institute, Brompton Hospital, London (MS); Department of Mathematics, Strathc1yde University, Glasgow (JM). Correspondence address: Dr. Alison Kerr, Monitoring Unit, Quarrier's Homes, Bridge of Weir, Renfrewshire, PAll 3SA, Scotland UK. The investigation was carried out at the Epilepsy Centre, Quar-
The Rett syndrome (RS) [1, 2] is associated with profound mental and physical disability in 1 in 10-12,000 girls [3,4] who may survive into the fourth decade. Signs appear in the first year of life [3-6] and exacerbate at 1-2 years. There are deficits in mental processing, extrapyramidal [3], pyramidal [7,8] and respiratory function. Diagnosis is easiest in the child between 5 and 15 years. rier's Homes, Bridge of Weir, Renfreshire, PAll 3SA in July 1987 and February 1988. Presented to British Paediatric Neurology Association in January 1988.
This is partly due to the occurrence of well defined episodes which include bursts of hyperventilation alternating with apnoeic pauses, exacerbations of stereotyped movements most marked in the face and limbs, non-seizure vacant spells and abnormal paroxysmal EEG activity.
mixer
ANV\
I-_:--:--:-l
f---+-
exp
timer tape
Resp
t--_ _ _--t'esplt'ace
97
02
o
In order to study these phenomena we constructed the apparatus shown in Fig 1 [9, 10] which records simultaneously the required parameters in the awake and unrestricted child. This is the second of two papers presenting our results. In our first paper [10] we described the main respiratory observations in 18 girls with RS (case numbers for that study are given in brackets in Table 1). Ten girls (1-10 of that paper) hyperventilated only when awake with development of severe hypocapnia. Hyperventilation was interspersed with prolonged apnoeic pauses (> 19 secs) with hypoxia in 47% of pauses. A further four patients who had hyperventilated in the past no longer hyperventilated but had frequent apnoeic pauses (cases 11-14 of that paper). The remaining four cases showed only occasional apnoeic pauses with Valsalva manoeuvres and were thought never to have hyperventilated. The aim of the present study was to establish the relationships between the non-seizure episodic events in EEG breathing and movement in 14 of the same girls and to observe changes with age.
PATIENTS AND METHODS Patients are listed in Table 1 with the essential respiratory
Fig 1 Method for recording behaviour. respiration and EEG.
Table 1 Fourteen Rett syndrome cases: respiratory. EEG and movement data Case number
1*5 (9)
2 (1)
Age (years)
6
6
Hypervenfilation present
+
+
(
*1 )
Lowest CO. recorded *2
3*6 (10) 6 +
4 (5)
5 (7)
6 (4)
7 (3)
8 (2)
9 (8)
10 (6)
7
7
10
11
12
13
16
+
+
+
+
+
+
+
1.8/18 2.2/24 2/13 1.4/13 1.6/8 2.1/18
11 (11 )
17 _*9
4.2/33 2.2/26 2.4/26 2.2/20 3.8/NR
]2*7 13*8 (15)
(18)
14 (16)
6
6
7
4.5/44 4/37 4.4/46
Apnoeic pauses present*3
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Lowest oxygen recorded*4
<50
60
<50
<50
70
<50
92
<50
92
66
80
94
98
97
EEG paroxysms at 1%-4 Hz
+
+
+
+
+
Increased in normal CO 2
+
+
+
+
+
4-6 Hz EEG activity
+ UR
+ UR
+ UR
+ R
+ UR
Paroxysmal 4-6 Hz activity Paroxysmal limb movement with resp. dysrhythmia
+
+
+
+
+
+ UR +*10
+ R
+ UR +
+
+ UR
+ UR
+ UR
+
+
+ UR
+
+
*1) Brackets after case numbers give case numbers for our fIrst paper. *2) Figure above the line is end-tidal carbon dioxide in volumes % (normal> 4); below the line is transcutaneous carbon dioxide mmHg (normal> 35 mm). * 3) Active expiratory apnoeic pauses, all included Valsa1va-like manoeuvers. *4) % saturation (normal> 97%). *5, *6, *7, *8) Cases with minor atypical features; *5) Better than expected use of hands. *6) Threatened miscarriage in pregnancy. *7 and * 8) Earlier stereotypic hand movements had almost disappeared. *9) Hyperventilation recorded in the past. *10) Abnormality increased when end-tidal CO. was normal. UR: unreactive EEG, R: reactive EEG. Anticonvulsant drugs: Carbamazepine (cases 2,3,5,8,9,11,12), sodium valproate (cases 6, 14), c10nazepam (cases 5, 13). NR: not recorded.
62 Brain & Development. Vol 12, No 1, 1990
data from our previous paper and the results of the present study. All girls were diagnosed as having RS and no other disorder. Minor atypical features in four girls are described in the footnote of the Table. Four girls included in our first paper were excluded because EEG was not simultaneously recorded. Twelve healthy girls, sisters and friends of RS patients were also studied (age range 6-18, median 14). Monitoring equipment Respiratory equipment: Full details are given in our previous paper [10] non-invasive respiratory monitors measured transcutaneous pC0 2 (Hewlitt Packard) end-tidal CO 2 (Engstrom) breathing movements by inductance plethysmography (respitrace, Studley data systems) and arterial oxygen saturation (Sa02) by pulse oximeter (Nellcor, modified to provide a beat to beat Sa0 2). The plethysmographic waveforms from the pulse oximeter were monitored to verify Sa02' All end-tidal CO 2 values were required to have adequate end-expiratory plateaus. Electroencephalographic recording: Seven bi-polar channels of EEG and one of ECG were recorded on a battery operated tape recorder using collodion applied scalp electrodes and chest electrodes with pre-amplifiers on a collar round the neck (medilog 9000, Oxford Medical Systems). EEG electrodes were applied in a modified montage in order to provide maximum information, bearing in mind the range of the child's activity as well as the need to correlate with other parameters. The EEG and ECG signals were relayed through a monitor/writer/coupling unit to a conventional EEG machine (Nihon Kohden). In all EEG recordings sensitivity was 7.5 microV per mm, time constant 0.3 secs with high frequency filter at 70 Hz. Electrode impedance was maintained between 2-5 Kohms. Video/audio recording: A portable camera (Sony video 8) was used for all recordings. A video mixer (ForA FVG 600) transmitted the EEG and the video picture of the subject, with seconds time trace, to a monitor screen and video recorder. Respiratory data was scanned directly from the individual respiratory monitors which were stacked beside the subject. Respiratory data was also stored independently on audio tape. Girls were monitored in daytime for periods of one to six hours (median 3). Seven RS girls slept for some part of the recording. In girls with severe hypocapnia due to hyperventilation, the carbon dioxide level was raised towards normal values (end-tidal 4.5 vol %, TcPC0 2 35 mmHg) by rebreathing inside a perspex hood or a mask. This did not cause distress and base line arterial oxygen saturations remained normal (> 97%). The twelve healthy girls, similarly recorded, were asked to hyperventilate in air and on separate occasions in the rebreathing hood.
Data extraction from records Abnormal episodes in respiration, EEG and movement were identified from video records. Such episodes were required to be distinctive and contain the same sequence of abnormal events on each occurrence. Episodes were defined independently for each girl. Episodes in behaviour were sequences of skeletal movement affecting trunk, face and limbs, identical on repeated occasions. Levels of this activity were defined for each girl (++, + or -). In addition to episodes of movement a record was made of interactions involving the girl. Respiratory episodes were defined as follows: 1) Hyperventilation: increase in respiratory excursion with subsequent fall in transcutaneous and end expiratory CO 2 level). 2) Apnoeic pause/s: pause or succession of pauses in respiratory excursion. 3) Valsalva-like manoeuvre: straining vocalisation in the course of apnoeic pause/so 4) Normal breathing: regular chest movement with rising CO 2 levels. "Other breaths" included irregular breaths other than those defined. Respiratory dysrhythmia is the term used to describe periods consisting of alternating episodes of hyperventilation and apnoeic pauses. Blood gas levels were confirmed from audio tape recordings. Electroencephalograms were analysed visually with particular reference to the distribution and responsiveness of rhythmic and/or paroxysmal activity and the presence or absence of paroxysmal slowing and epileptogenic activity. Non epileptic 'paroxysmal activity was characterised by sudden increase in amplitude of slow components (delta and theta) lasting more than 2 secs. The state of each patient was recorded as asleep, awake or transitional. Episodes from each area of function were viewed "blind" by covering the parts of monitor screen and chart showing other areas of function. EEG records and respiratory records were reviewed independently by individual authors. All the data from each patient's video tape was charted on continuous paper which was divided into columns for each area of function in which episodes and states were to be correlated. Time was marked at one horizontal line for 4 seconds. Onset and end of episodes were recorded in seconds. In cases where episodes were sharply defined, the numbers and durations of each type of episode and state and the associations between them were calculated. Other records were reported in the conventional manner with events in other parameters being taken into account in the reporting of each.
RESULTS Patients fall into 3 groups according to results (Table 1).
Kerr et al: EEG respiration and movement in Rett syndrome 63
Group 1 (Six girls, cases 1-6,6-10 years)
All these girls had episodes of severe hyperventilation with hypocapnia and alternating apnoeic pauses and all showed non-epileptic paroxysmal slow activity on EEG. These paroxysmal EEG changes were generally asso· ciated with periods of normal breadling (and normal CO 2 levels) whether the girls were alert, drowsy or in day·time sleep . They seldom occurred during episodes of hyperventilation. In cases 1-4 the paroxysmal EEG change consisted of 1~-4 Hz activity of 75-250 microV in amplitude. In case 5 it consisted of slow spike and wave discharges lasting 10-30 sees and too-I 50 microV in amplitude. In case 6 the change consisted of paroxys· mal bursts of 4-6 Hz activity of 50-75 microV amplitude. Fig 2 illustrates the relatively "normal" EEG seen during HV and the increase in abnormality which occurred during normal awake breathing in cases 4 and 5. Fig 3 shows the parallel abnormalities in EEG and respiration
in case 5. For cases 1, 2, 4 and 5 in whom episodes were most clearly demarcated, seizure-free periods were selected to allow comparison beween the frequency of these episodes in periods of dysrhythmic breathing, normal breathing (alert state) and normal breathing (drowsy or asleep). During periods of respiratory dysrhythmia, hypocapnia was always present. Within these selected periods significant hypoxia < 50% was recorded on one occasion without alteration in EEG or behaviour. Table 2 compares the time occupied by the nonepileptic EEG paroxysms during periods of established normal and dysrhythmic breathing during the selected periods in these fou r girls. Chi-squared tests were carried out in cases I, 2 and 5 to compare regular discrete EEG paroxysms in the various periods of breathing. There was a clear difference in the number of buTSis during respiratory dysrhythmia (Jow) and normal breathing
•
• • ••
,~
TC
'----------------------------------------I~·-----------------------------------
0.> _
~ .~
7O.vL-
v..__t" ...:.,_._.!~~
~jv:Jl..·p.;A.--!;t..~.IWA-..-Jt,.'.I\,.-J~ {~L,.J.".,..1r!.r. .~/'-·~,.
'-
~V~'V'C/_ ' \J,""''-,v\i\A/'vv-.'u\ir;\.-\rL;''.~v7 ~/""f\-NI..rl'·;v.r.Nf'>JV'c.l'-I'
,
-~~~V""""'-~-
~~~ .r'""r'v~;II.,'i\I'r.J" f,!,"__f'r't,('iY/Vy ·~/~MV'__('.rvv..rvf'·Iv-. . ,...,.r-..~/\i\/V. ..r_v
,~
Tt
0.3_
• ••
• ~\.----~~.,.,r..-..
CH HOC!
~~Iv-.~/.-'. A...IvI'A--A--~~A.-..~~.A~vA~V-----~~--~--~~~~~v~
~~----~~~~~~~~--~~~~
•
,~
,,~
l .C. u _
• •• ~,~
•
"""
L
IO.VL -
,
'-
•
I
t'~.
., I
Fig 2 Electroencephalograms in cases of Rett ryndrome. EXQmpJes sho w the EEG in cares 4 ( ) arul j ( oj and c); and dun'ng normal breathing b) and d). Example e) sho ws the theta rhythm in case 10 ( j.
64 Brain
&;
Development, Vol 12. No 1,1 990
)
during hyperventilation
JTTTITTTl""'~'TTTTTTTT'J"TTTTTTTTLJ'TTTTTTTT'~~
~V\')~\I
i~~j~~Jr~~
I
i~~"'~'~lr~
~""j~,.r""~ i
r~
~
. "
f" "
r. ~ f
, I . . . ,. ~ ... } ,~
rr r ~
,
~
, ( f '.• t ... ' !
~
r
~
i1,
.. ' , I
f
Fig 3 Parallel records of EEG breathing and behaviour in case 5. 1, 2 and 3 recordings during a period of respiratory dysrhythmia. 4, 5 and 6 recordings during alert normal breathing (while rebreathing). Small letters indicate respiratory activity corresponding to the photographs. In case 5 the paroxysms of movement ++ occurred not only during periods of respiratory dysrhythmia but specifically during the HVepisodes within these periods. During normal breathing, movements diminished.
(high), p = 0.001 for this difference. The number of bursts occurring in alert and asleep periods of normal breathing was not significantly different. In cases 1, 2, 4 and 5 the EEG paroxysms occupied 1-3% of the time during respiratory dysrhythmic periods and 8-100% during alert normal breathing periods (p~O.OOl for this difference). Episodes of stereotyped movement involved face trunk limbs and hands, and were graded ++ when they were very energetic with wringing or clapping of the hands, + when they were intermediate with tense posturing of the limbs and - when absent or minimal with fingertip movement. In cases 1, 2, 4, 5 and 6 episodes of stereotyped move-
ment (++ and +) occurred during the periods of respiratory dysrhythmia and diminished (+ and -) during periods of normal breathing (see Table 2). The percentages of ++ activity for the nineteen continuous periods recorded were compared using a non parametric test (KruskalWallis) which assumes that the various periods are independent. This test shows a significant difference in amount of ++ movement between dysrhythmic and normal alert breathing periods (p = 0.01). In case 5 the episodes of vigorous movement (++) coin. cided precisely with the episodes of hyperventilation and during apnoeic pauses reduced instantaneously to the less intense level (+), so that limb and chest movements oc-
Kerr et al: EEG respiration and movement in Rett syndrome 65
Table 2 Comparison of non-epileptic EEG paroxysms occurring during periods of normal and dysrhythmic breathing in four cases
Case
Time sec
ETC0 2 vol%
no
EEG bursts %o/time
Respiratory dysrhythmia: 741 (2) 1.9-3 1 1 2.8-3.5 1,784 (4) 14 2 2-3 2 1,108 (3) 4 2-3 1,264 (3) 5 Normal breathing: alert 4-5 124 (2) 5 1 4.9 239 (1) 9 2 3.8-4 389 (2) 4 Contino 3.1-4.4 985 (2) 36 5 Normal breathing: asleep or drowsy 4-5 448 1 25 4-5 623 2 40 4 3.7-5 Contino 597 4.7-4.9 150 5 9
Movement % o/time
++
<1 3 2 <1
92 38 63 54
8 8 100 57
0 0 0 0
+
3 56 35 46
5 6 3 0
48 52 22 78 32 68 98 2
16 24 61 29
In case 5 movement ++ coincided with HV and + with apnoeic pauses. ( ) Brackets indicate the numbers of periods analysed.
curred exactly together. Fig 3 illustrates these events. An agitated manner accompanied stereotyped movement ++ and a preoccupied expression appeared during Valsalva-like manoeuvers. In case 3 the stereotyped activity was not episodic. The change from a period of normal breathing to respriratory dysrhythmia was seen when a girl woke spontaneously, was awakened, was offered an object which captured her interest, or was spoken to directly. At these' times the episodes of respiratory and limb dysrhythmia commenced simultaneously giving the impression of a burst of movement outwith her control but linked to her alerting. The change from periods of respiratory dysrhythmia to normal breathing was recorded when a girl became drowsy, following a period of hiccoughs or crying and when rebreathing had raised the level of inspired CO 2 towards normal. Group 2 (Five girls, cases 7-11, aged 11-17 years) These were the remaining girls with severe respiratory dysrhythmia. In cases 7 to 10 this included hyperventilation with hypocapnia and in case 11 only the apnoeic pauses with Valsalva-like manoeuvers. In these girls the regular stereotyped limb activity was not paroxysmal. Valsalva-like manoeuvers were accompanied by an anxious appearance. On EEG all five girls showed, as well as some polymorphic delta activity, prominent 4-6 Hz theta activity which was often synchronous and generalised. This was largely unreactive to normal physiological stimuli but did
66 Brain & Development, Vol 12, No 1, 1990
show some variation in amount and amplitude from time to time. In cases 8 and 11 the 4-6 Hz activity showed paroxysmal increase in amplitude towards the end of apnoeic pauses. Fig 2 illustrates the persistent rhythmic 4-6 Hz activity in case 10. Group 3 (Three girls, cases 12-14, aged 6-7 years) These girls showed only occasional apnoeic pauses with Valsalva-like manoeuvers and normal gases and had no history of hyperventilation. Their stereotyped limb movements were not episodic. Case 14 had paroxysmal slow activity at 1~-4 Hz (in association with spike discharges) which fluctuated in amount and was particularly prominent during sleep. Seizures and non-seizure vacant spells Six children (cases 1, 2, 3,4, 5 and 6) had attacks of vacancy and staring not associated with significant EEG changes. Cases 1 and 2 each also had a minor complex seizure, with appropriate ictal activity on EEG during attacks. Both the non-seizure spells and the seizures occurred after periods of intense hyperventilation with severe hypocapnia and were associated with apnoeic pauses with or without a Valsalva-like manoeuver. Pulse oximetry indicated subsequent but not prior hypoxaemia. Nine of the 14 RS girls (cases 1, 2, 5, 6, 7, 8, 9, 12 and 14) showed some epileptogenic activity in the form of single or grouped spike and sharp wave discharges or slow spike and wave, distinct from the episodes of paroxysmal slow activity. Our data on seizure and non-seizure vacant spells will be the subject of a separate publication. All the RS girls showed general EEG abnormality with excess of polymorphic slowing and poorly developed (daytime) sleep change. Case 4 showed least abnormality although the EEG was poorly developed. The 12 healthy girls hyperventilated voluntarily for 3.9-15.8 minutes (median 7.8). Carbon dioxide levels dropped to 1.9-3.3 vol% (median 2.1). Nine (all those under 15) showed a normal slow wave response to this hypocapnia. The three who failed to show a normal response were aged 18, 15 and 18 years and developed CO 2 levels of 3.3,2.4 and 1.9 vols% respectively. None of the girls showed limb or face movement comparable to that of the RS girls.
DISCUSSION The striking observations from this study are the association of episodic non-epileptic paroxysmal abnormality on EEG with periods of normal breathing; and the association of the episodes of repetitive movement with the respiratory dysrhythmia. These are so consistent in the younger girls as to provide useful diagnostic signs. As with other features of RS these seem related to a particular age period between about 4 years (following
regression) and the early teens, after which dysrhythmic breathing is less inclined to include hyperventilation with hypocapnia and both the limb movements and EEG changes may become less episodic. The occurrence in these very disabled girls, of such complex stereotyped paroxysms of movement involving face trunk and limbs simultaneous with the respiratory dysrhythmia seems to suggest a common trigger for both which would appear to be outwith the voluntary control of the child although possibly linked to alerting. In cases 1, 2, 4 and 5 it is notable that the episodic non-epileptic paroxysmal EEG activity occurred when breathing and respiratory gases were normal, whether after waking from normal sleep, during daytime sleep, during rebreathing or at other times, while on the other hand, this paroxysmal activity was seldom seen during hypocapnia and none of the RS girls showed a normal slow wave response to hyperventilation. In normal children, particularly between 6 and 12 years slow wave activity develops during hyperventilation, associated with hypocapnia [11]. Absence of this response is not abnormal but would be unusual with significant hypocapnia, especially in a younger age group. This normal slow wave response was seen in 9 of our 12 healthy girls including all the youngest. The mechanisms underlying normal EEG changes during hyperventilation with hypocapnia are poorly understood in spite of a good deal of work having been done over many years. The subject has been carefully reviewed by Patel and Morsby [12] although they note that their conclusions are speCUlative. We cannot equate the paroxysmal abnormality of the RS girls with the normal EEG response to hyperventilation since the amplitudes and wave forms are different. Our findings therefore remain to be explained. The nature of these changes and their presence during drowsy and asleep states as well as in the apparently alert child might indicate that they reflect the existence of a pathological "non-alert" state, a consequence of the essential developmental deviation responsible for the syndrome. Such a state might be expected to provoke an extreme alerting reaction. The prominent 4-6 Hz activity is an interesting feature of the RS cases. This has been described [13-15] but has been considered a feature of slowing with increasing age. On the contrary it was present in our youngest girls, appeared reactive in case 4, the child with the least abnormal record and in case 7, while in the other cases it was unreactive to physiological stimuli and behavioural change although fluctuating from time to time. In four cases [1,6,8,11] it showed occasional paroxysmal increase in amplitude. This theta activity which is a normal feature of childhood appears in RS girls to develop as expected, then to become unresponsive and finally totally unreactive and persistent. It seems to be highly characteristic of these
girls. In our series slow activity in the delta range, whether polymorphic, rhythmic and/or paroxysmal, was more prominent in the younger age groups. In this paper we have not fully presented or discussed data on the important relationships between seizures and the respiratory dysrhythmia. However our observations of partial complex seizures and non-seizure vacant spells indicate that without EEG monitoring it is difficult to distinguish minor seizures from other attacks and that both kinds of attack may relate to the respiratory abnormalities. We now plan to explore further both seizure and nonseizure vacant spells, to investigate the respiratory control mechanisms in RS, awake and asleep and to observe the effects of morphine antagonists on the episodic phenomena which we have documented. It is no surprise that a pathological process producing profoundly disordered thinking and movement may also involve respiratory control. The assumption that RS may be due to a single gene defect makes the neural mechanisms underlying these disordered functions of peculiar interest.
ACKNOWLEDGMENETS The authors gratefully acknowledge the technical assistance of Mr. E. Picton Jones, Dr. A. Etchells, and staff of the medical illustration department RHSC also the support of medical colleagues and the families of subjects. Dr. Kerr is funded by Quarrier's Homes, the Rett Syndrome Associations (National and UK) and Dr. Southall and Dr. Samuels by the National Heart and Chest Hospital. Equipment was funded by the Scottish Society for the Mentally Handicapped, National Rett Syndrome Association, Labaz-Sanofi, Oxford Medical and Nellcor.
REFERENCES 1. Rett A. Uber ein eigenartiges hirnatrophisches Syndrom bei Hyperammonamie im Kindsalter. Wien Med Wochenschr 1966;116:723-6. 2. The Rett Syndrome Diagnostic Criteria Work Group. Diagnostic criteria for Rett syndrome. Ann Neurol 1988;23: 425-8. 3. Kerr A, Stephenson J. A study of the natural history of Rett syndrome in 23 girls. Am J Med Genet 1986;24(suppll): 77-83. 4. Hagberg B. Rett syndrome: Swedish approach to analysis of prevalence and cause. Brain Dev (Tokyo) 1985;7:277-80. 5. Nomura Y, Segawa M, Hasegawa M. Rett syndrome - clinical studies and pathophysiological consideration. Brain Dev (Tokyo) 1984;6:475-86. 6. Kerr A, Montague J, Stephenson J. The hands and the mind pre- and post-regression in Rett syndrome. Brain Dev (Tokyo) 1987;9:487-90. 7. Eyre J, Kerr A, Miller S, O'Sullivan M, Ramesh V. Neurophysiological observations on the corticospinal projections to the upper limb in subjects with Rett syndrome. (submitted for publication) 8. Oldfors A, Hagberg B, Nordgren H, Sourander P, Witt-Enger-
Kerr et al: EEG respiration and movement in Rett syndrome 67
strom 1. Rett syndrome: spinal cord neuropathology. Pediatr Neural 1988;4:172-4. 9. Kerr A, Amos P, Etchells A, Irwin A, Holmes T, Stephenson J. A low cost method for simultaneous video recording of ambulant subject and electroencephalograph: the Quarrier's system. J Ment Defic Res 1988;32:497-500. 10. Southall D, Kerr A, Tirosh E, Amos P, Lang M, Stephenson J. Hyperventilation in the awake state: potentially treatable component of Rett syndrome. Arch Dis Child 1988;63: 1039-48. 11. Binnie C, Coles P, Margerison J. The influence of end-tidal carbon dioxide tension on EEG changes during routine hyperventilation in different age groups. Electraencephalagr Clin Neuraphysial 1969;27:304-6.
68 Brain & Development, Vol 12, No 1, 1990
12. Patel V, Maulsby R. How hyperventilation alters the electroencephalogram: a review of controversial viewpoints emphasizing neurophysiological mechanisms. J Clin NeuraphysiaI1987;2:101-20. 13. Niedermeyer E, Rett A, Renner H, Murphy M, Naidu S. Rett syndrome and the electroencephalogram. Am J Med Genet 1986;24(suppl1):195-9. 14. Verma N, Cheda R, Nigro M, Hart Z. Electroencephalographic findings in Rett syndrome. Electroencephalagr Clin Neuraphysial 1986 ;64: 394-40 1. 15. Hagne I, Witt-Engerstrom I, Hagberg H. EEG development in Rett syndrome. A study of 30 cases. Electraencephalagr Clin NeuraphysiaI1989;72:1-6.