October 1978
704
TheJournalofPEDIATRICS
Body rocking, head banging, and head rolling in normal children Body rocking, head banging, and head rolling-three rh)'thmic behaviors that involve sthmdation of the vestibular s)'stem-were studied in 525 nortnal chiMren. Data anal)'ses focused on two issues: (1) the prevalence and duration of these habits and their relationship to child and family factors and (2) the hypothesis that children who persistentl)' displa)'ed such behaviors would be reported as developmentall)" tnore advanced than "non-self-sthnttlators." Body rocking was the earliest to appear and most prevalent of the three habits. Head banging and head rolling had roughl)' the same prevalence and age of onset. Comparisons of "self-stimulators" with "'non-self-stbnulators" )ielded no significant effects for birth order or SES; onl)'for head banging was a significant (3:1) ratio of males to females found. A comparison of the ages at which 12 "'milestones" first appeared supported the h)pothesis of developtnental precocit)' for the body rockers and the herod bangers, but not for the head rollers.
Frank Sallustro, M.A., and Constance W. Atwell, Ph.D.,* C l a r e m o n t , Calif.
A L T H o u o l l MUCti ATTENTION has been given to
habitual, rhythmic behaviors among autistic and retarded children, little in-depth study has been made of the normal child. Of the three rhythmic habits-body rocking, head banging, and head rolling-only head banging in the normal child has attracted reasonable inquiry; body rocking and head rolling have remained virtually uninvestigated. With the exception of incidence figures," which vary from 3.3%' to 15.2%,"- some findings about head banging seem well confirmed. Both de Lissovoy~and Kravitz et al' agreed that (1) head banging begins typically at 8.6 months and stops usually before the fourth year; (2) the ratio of head banging for males and females is approximately 3 to I; (3) prior to the onset of their head banging, most head bangers display other rhythmic habits, most notably body rocking; and (4) head banging takes place most often at bedtime. Other studies have found birth order,:- '- ~ low socioeconomic status, 2 high responsiveness to music, and excellence in motor coordination ~ to be important correlates of head banging, but the value of From Claremont Gra&tate School and Pitzer College. Supported b)' grants froth the 1t711iam T. Grant Foundation and HD09644 to C. II:A. Presented in part to the Western Psychological Association, Seattle. Washhlgton, April 23, 1977. *Reprint address: Pitzer College, Clarethoht. CA 91711. 1Iol. 93, No. 4, pp. 704-708
these findings has been undermined by the lack of suitable control groups. Nevertheless, the evidence afforded by these studies suggests that head bangers may be quite different from non-head bangers on several dimensions. Abbreviation used SES: socioeconomic level
]
The only firm figure on habitual rockers came from de I_issovoy-": approximately 21% of his sample of 374 children had high levels of this behavior. Lourie ~observed that, on the average, children began rocking between 6 and 10 months of age and stopped by 289 to 3 years; Kravitz and Boehm 7 found a median onset of 6.1 months in their month-by-month study of 200 normal infants. For head rolling behavior, Kravitz and Boehm reported a 10% incidence, the median age of onset be!ng 12+ months. This 10"7o figure is, however, a likely overestimate of habitual heaa rollers, since Kravitz and Boehm focused on the initial appearance of head rolling, not its persistence. The importance of studying such rhythmic self-stimulatory behaviors is emphasized by the growing body of literature on the effects of externally imposed movement (vestibular) stimulation on behavioral development. Thoman and Korner s showed that vestibular stimulation was significantly more effective than tactile stimulation in 0022-3476/78/100704+05S00.50/0 9 1978 The C. V. Mosby Co.
Vohmie 93 Number 4 reducing the crying time of newborn rats. In another study, the superiority of vestibular stimulation over contact stimulation in reducing arousal level (crying time) was again obtained? Ter Vrught and Pederson '~ reported that general arousal was most reduced in human infants when vestibular stimulation (vertical rocking) was given at certain "peak" levels of frequency and amplitude. Apart from reducing arousal, vestibular stimulation appears effective in evoking visual alertness 8. " and visual tracking,'"- and in enhancing weight gains, '~ motor responsiveness," and neurologic development '~ in premature infants. Rice's '~ report showed that premature infants, given enriched tactile and vestibtilar stimulation one hour per day for 30 days, surpassed not only other premature infants in neurologic development, but normal, term infants as well. In order to resolve the inconsistencies in the previously reported incidence and demographic data for rhythmic habits, and to assess the effects of large amounts of self-induced vestibular stimulation within the same sample on which the incidence figures were collected, a study was undertaken of clinically normal, nonhospitalized children who persistentl), engage in body rocking, head banging, or head rolling. For each of these rhythmic habits, the following information has been obtained by retrospective questionnaires: (1) incidence in an unselected pediatric population; (2) onset and duration; (3) relationship of rhythmic habit to child and family characteristics, such as sex, birth order, race, and socioeconomic status; and (4) situations in which the rhythmic habit occurs. Additionally, habitual self-stimulators were compared with non-self-stimulators on 12 measures of motor development; the hypothesis was advanced that "vestibular self-stimulators" would have reached developmental "milestones" earlier than would non-self-stimulators. METHOD Subjects. The subjects were 525 healthy children (256 girls and 269 boys) who were brought to the Kaiser Pediatric Clinic in Fontana, California, for their regular medical examination. Their ages ranged from three months to six years (median age: 18 months). The subject population represents a wide range of ethnic, geographic (urban-rural), and socioeconomic level diversity. Two measures of SES were used. One measure was based on the "head of household's" occupation, which, in almost all cases, was the father's. Only when the father was absent or unemployed was the mother's occupation used. T'he second SES measure considered only those families in which both parents worked. When this occurred-it was uncommon, with'less than 30% of the families having two working parents-an average occupation rating was
Body rocking, head banging, and head rolling
705
Table !. Responses to questions regarding when body rocking, head banging, and head rolling occurred
Percentage of yes responses "Questions tired or at bedtime When awakening When alone in crib or playpen When upset or irritable When listening to music When riding in a car When
Body I llead rockers bangers
I
tlead rollers
55.8 46.8 67.7
66.7 20.8 52.4
56.5 21.7 81.0
25.0 79.2 32.9
56.0 24.0 20.0
28.0 50.0 23.1
obtained. Both measures were derived from the work of Blau and Duncan. '~ Distribution by ethnic group was: white (71%), Mexican-American (21%), black (6%), and Oriental (1%). Administration of questionnaires. When each child arrived at the clinic, a receptionist gave the caretaker a 78-item questionnaire to complete before leaving. To avoid any systematic bias, all receptionists were asked to give every caretaker a questionnaire immediately upon arrival. Although this was not always done, it was done routinely enough to suggest that this sample was an unselected one. From September, 1975, to June, 1976, 585 questionnaires were distributed. Of that number, 525 (90'70) were returned fully completed or almost fully completed, though the N varied considerably for individual questions. Behavioral criteria. Only those children who were judged by their caretakers as having engaged in one or more rhythmic behaviors persistentl), and intensely qualified as self-stimulators. Thus, in reply to the question, "Has'this child ever engaged in repetitive head banging (other than during a tantrum)?" only those coded "3" (a few times a week), "4" (once a day), or "5" (more than once a day) were considered "head bangers." Those coded "2" (rarely) were combined with those coded "1" (never) to form the "non-head bangers." The same procedure was followed for determining the "head-rollers." In response to the question, "Has this child ever engaged in body rocking while on his hands and knees?" a different scoring haethod was used. Only those coded "4" (once a day) or "5" (more than once a day) were termed "body rockers"; codes "1" and "2" comprised the "nonrockers." Those coded "3" (a few times a week) were excluded from all analyses, as well as the 51 children who the caretaker indicated rocked when learning to crawl, but not under other circumstances. Developmental variables. To determine whether selfstimulators were developmentally more advanced than
706
Salhtstro and A twell
The Journal of Pediatrics October 1978
Table !I. Comparisons of mean ages (in months) when developmental milestones first appeared tlabitual rockers Milestone
Lifted head from the mattress Started playing with hands and feet tleld head erect Reached for a toy or crib gym Rolled from stomach to back Rolled from back to stomach Sat up without support Got up onto hands and knees Began to crawl Pulled up to a standing position Began to "cruise" (holding on to furniture) Walked without support
Mean (N)
Non -rockers
SD
Mean (N)
SD
t vahte*
1.03 (80) 2.51 (71). 2.69 (67) 3.01 (73) 3.35 (74) 3.90 (73) 5.47 (76) 5.70 (76) 6.06 (78) 7.40 (77) 8.14 (76)
i.I 1 1.07 1.41 !.10 1.38 1.48 1.26 1.48 1.48 !.56 1.71
i.33 (224) 2.75 (204) 2.80 (I 94) 3.23 (194) 3.4 ! (200) 4.07 (189) 5.63 (216) 6.31 (198) 6.59 (217) 7.88 (224) 8.91 (229)
1.19 1.18 1.45 1.14 1.30 1.44 1.35 1.56 1.50 1.68 1.79
- i.96~ -1.51 -0.54 -1.42 -0.32 -0.85 -0.90 -2.92w -2.69w -2.20~ -3.29w
!.41 0.86 0.00 i.20 1.29 2.30 3.89:1: 10.30w 4.67~ 2.70 7.31w
10.71 (42)
2.20
11.07 (203)
i .97
- i .06
3.94:1:
Adjusted Ft
*Subject's age was not removed. *Subject's age was removed as a covariate. ~P < 0.05. w
704
TheJournalofPEDIATRICS
Body rocking, head banging, and head rolling in normal children Body rocking, head banging, and head rolling-three rh)'thmic behaviors that involve sthmdation of the vestibular s)'stem-were studied in 525 nortnal chiMren. Data anal)'ses focused on two issues: (1) the prevalence and duration of these habits and their relationship to child and family factors and (2) the hypothesis that children who persistentl)' displa)'ed such behaviors would be reported as developmentall)" tnore advanced than "non-self-sthnttlators." Body rocking was the earliest to appear and most prevalent of the three habits. Head banging and head rolling had roughl)' the same prevalence and age of onset. Comparisons of "self-stimulators" with "'non-self-stbnulators" )ielded no significant effects for birth order or SES; onl)'for head banging was a significant (3:1) ratio of males to females found. A comparison of the ages at which 12 "'milestones" first appeared supported the h)pothesis of developtnental precocit)' for the body rockers and the herod bangers, but not for the head rollers.
Frank Sallustro, M.A., and Constance W. Atwell, Ph.D.,* C l a r e m o n t , Calif.
A L T H o u o l l MUCti ATTENTION has been given to
habitual, rhythmic behaviors among autistic and retarded children, little in-depth study has been made of the normal child. Of the three rhythmic habits-body rocking, head banging, and head rolling-only head banging in the normal child has attracted reasonable inquiry; body rocking and head rolling have remained virtually uninvestigated. With the exception of incidence figures," which vary from 3.3%' to 15.2%,"- some findings about head banging seem well confirmed. Both de Lissovoy~and Kravitz et al' agreed that (1) head banging begins typically at 8.6 months and stops usually before the fourth year; (2) the ratio of head banging for males and females is approximately 3 to I; (3) prior to the onset of their head banging, most head bangers display other rhythmic habits, most notably body rocking; and (4) head banging takes place most often at bedtime. Other studies have found birth order,:- '- ~ low socioeconomic status, 2 high responsiveness to music, and excellence in motor coordination ~ to be important correlates of head banging, but the value of From Claremont Gra&tate School and Pitzer College. Supported b)' grants froth the 1t711iam T. Grant Foundation and HD09644 to C. II:A. Presented in part to the Western Psychological Association, Seattle. Washhlgton, April 23, 1977. *Reprint address: Pitzer College, Clarethoht. CA 91711. 1Iol. 93, No. 4, pp. 704-708
these findings has been undermined by the lack of suitable control groups. Nevertheless, the evidence afforded by these studies suggests that head bangers may be quite different from non-head bangers on several dimensions. Abbreviation used SES: socioeconomic level
]
The only firm figure on habitual rockers came from de I_issovoy-": approximately 21% of his sample of 374 children had high levels of this behavior. Lourie ~observed that, on the average, children began rocking between 6 and 10 months of age and stopped by 289 to 3 years; Kravitz and Boehm 7 found a median onset of 6.1 months in their month-by-month study of 200 normal infants. For head rolling behavior, Kravitz and Boehm reported a 10% incidence, the median age of onset be!ng 12+ months. This 10"7o figure is, however, a likely overestimate of habitual heaa rollers, since Kravitz and Boehm focused on the initial appearance of head rolling, not its persistence. The importance of studying such rhythmic self-stimulatory behaviors is emphasized by the growing body of literature on the effects of externally imposed movement (vestibular) stimulation on behavioral development. Thoman and Korner s showed that vestibular stimulation was significantly more effective than tactile stimulation in 0022-3476/78/100704+05S00.50/0 9 1978 The C. V. Mosby Co.
Vohmie 93 Number 4 reducing the crying time of newborn rats. In another study, the superiority of vestibular stimulation over contact stimulation in reducing arousal level (crying time) was again obtained? Ter Vrught and Pederson '~ reported that general arousal was most reduced in human infants when vestibular stimulation (vertical rocking) was given at certain "peak" levels of frequency and amplitude. Apart from reducing arousal, vestibular stimulation appears effective in evoking visual alertness 8. " and visual tracking,'"- and in enhancing weight gains, '~ motor responsiveness," and neurologic development '~ in premature infants. Rice's '~ report showed that premature infants, given enriched tactile and vestibtilar stimulation one hour per day for 30 days, surpassed not only other premature infants in neurologic development, but normal, term infants as well. In order to resolve the inconsistencies in the previously reported incidence and demographic data for rhythmic habits, and to assess the effects of large amounts of self-induced vestibular stimulation within the same sample on which the incidence figures were collected, a study was undertaken of clinically normal, nonhospitalized children who persistentl), engage in body rocking, head banging, or head rolling. For each of these rhythmic habits, the following information has been obtained by retrospective questionnaires: (1) incidence in an unselected pediatric population; (2) onset and duration; (3) relationship of rhythmic habit to child and family characteristics, such as sex, birth order, race, and socioeconomic status; and (4) situations in which the rhythmic habit occurs. Additionally, habitual self-stimulators were compared with non-self-stimulators on 12 measures of motor development; the hypothesis was advanced that "vestibular self-stimulators" would have reached developmental "milestones" earlier than would non-self-stimulators. METHOD Subjects. The subjects were 525 healthy children (256 girls and 269 boys) who were brought to the Kaiser Pediatric Clinic in Fontana, California, for their regular medical examination. Their ages ranged from three months to six years (median age: 18 months). The subject population represents a wide range of ethnic, geographic (urban-rural), and socioeconomic level diversity. Two measures of SES were used. One measure was based on the "head of household's" occupation, which, in almost all cases, was the father's. Only when the father was absent or unemployed was the mother's occupation used. T'he second SES measure considered only those families in which both parents worked. When this occurred-it was uncommon, with'less than 30% of the families having two working parents-an average occupation rating was
Body rocking, head banging, and head rolling
705
Table !. Responses to questions regarding when body rocking, head banging, and head rolling occurred
Percentage of yes responses "Questions tired or at bedtime When awakening When alone in crib or playpen When upset or irritable When listening to music When riding in a car When
Body I llead rockers bangers
I
tlead rollers
55.8 46.8 67.7
66.7 20.8 52.4
56.5 21.7 81.0
25.0 79.2 32.9
56.0 24.0 20.0
28.0 50.0 23.1
obtained. Both measures were derived from the work of Blau and Duncan. '~ Distribution by ethnic group was: white (71%), Mexican-American (21%), black (6%), and Oriental (1%). Administration of questionnaires. When each child arrived at the clinic, a receptionist gave the caretaker a 78-item questionnaire to complete before leaving. To avoid any systematic bias, all receptionists were asked to give every caretaker a questionnaire immediately upon arrival. Although this was not always done, it was done routinely enough to suggest that this sample was an unselected one. From September, 1975, to June, 1976, 585 questionnaires were distributed. Of that number, 525 (90'70) were returned fully completed or almost fully completed, though the N varied considerably for individual questions. Behavioral criteria. Only those children who were judged by their caretakers as having engaged in one or more rhythmic behaviors persistentl), and intensely qualified as self-stimulators. Thus, in reply to the question, "Has'this child ever engaged in repetitive head banging (other than during a tantrum)?" only those coded "3" (a few times a week), "4" (once a day), or "5" (more than once a day) were considered "head bangers." Those coded "2" (rarely) were combined with those coded "1" (never) to form the "non-head bangers." The same procedure was followed for determining the "head-rollers." In response to the question, "Has this child ever engaged in body rocking while on his hands and knees?" a different scoring haethod was used. Only those coded "4" (once a day) or "5" (more than once a day) were termed "body rockers"; codes "1" and "2" comprised the "nonrockers." Those coded "3" (a few times a week) were excluded from all analyses, as well as the 51 children who the caretaker indicated rocked when learning to crawl, but not under other circumstances. Developmental variables. To determine whether selfstimulators were developmentally more advanced than
706
Salhtstro and A twell
The Journal of Pediatrics October 1978
Table !I. Comparisons of mean ages (in months) when developmental milestones first appeared tlabitual rockers Milestone
Lifted head from the mattress Started playing with hands and feet tleld head erect Reached for a toy or crib gym Rolled from stomach to back Rolled from back to stomach Sat up without support Got up onto hands and knees Began to crawl Pulled up to a standing position Began to "cruise" (holding on to furniture) Walked without support
Mean (N)
Non -rockers
SD
Mean (N)
SD
t vahte*
1.03 (80) 2.51 (71). 2.69 (67) 3.01 (73) 3.35 (74) 3.90 (73) 5.47 (76) 5.70 (76) 6.06 (78) 7.40 (77) 8.14 (76)
i.I 1 1.07 1.41 !.10 1.38 1.48 1.26 1.48 1.48 !.56 1.71
i.33 (224) 2.75 (204) 2.80 (I 94) 3.23 (194) 3.4 ! (200) 4.07 (189) 5.63 (216) 6.31 (198) 6.59 (217) 7.88 (224) 8.91 (229)
1.19 1.18 1.45 1.14 1.30 1.44 1.35 1.56 1.50 1.68 1.79
- i.96~ -1.51 -0.54 -1.42 -0.32 -0.85 -0.90 -2.92w -2.69w -2.20~ -3.29w
!.41 0.86 0.00 i.20 1.29 2.30 3.89:1: 10.30w 4.67~ 2.70 7.31w
10.71 (42)
2.20
11.07 (203)
i .97
- i .06
3.94:1:
Adjusted Ft
*Subject's age was not removed. *Subject's age was removed as a covariate. ~P < 0.05. w
