Neurobehavioral functioning and magnetic resonance imaging findings in young boxers

Full access

✓ In a prospective investigation of neurobehavioral functioning in young boxers, 13 pugilists and 13 matched control subjects underwent tests of attention, information-processing rate, memory, and visuomotor coordination and speed. The results disclosed more proficient verbal learning in the control subjects, whereas delayed recall and other measurements of memory did not differ between the two groups. Reaction time was faster in the boxers than in the control subjects, but no other differences were significant. Ten subjects in each group were retested 6 months later and exhibited improvement in their neuropsychological performance as compared to baseline measurements. However, there were no differences in scores between the boxers and the control subjects at the follow-up examination or in the magnitude of improvement from baseline values. Magnetic resonance imaging, which was performed in nine of the boxers, disclosed normal findings.

Abstract

✓ In a prospective investigation of neurobehavioral functioning in young boxers, 13 pugilists and 13 matched control subjects underwent tests of attention, information-processing rate, memory, and visuomotor coordination and speed. The results disclosed more proficient verbal learning in the control subjects, whereas delayed recall and other measurements of memory did not differ between the two groups. Reaction time was faster in the boxers than in the control subjects, but no other differences were significant. Ten subjects in each group were retested 6 months later and exhibited improvement in their neuropsychological performance as compared to baseline measurements. However, there were no differences in scores between the boxers and the control subjects at the follow-up examination or in the magnitude of improvement from baseline values. Magnetic resonance imaging, which was performed in nine of the boxers, disclosed normal findings.

Recent studies of the neurobehavioral sequelae of boxing5,23,25,37,38 have generated interest and concern about the health risks to participants and controversy over banning this sport.8,19,29,34 Computerized tomography (CT) has disclosed evidence of cerebral atrophy (including prominent sulci and ventricular enlargement) and/or of a cavum septum pellucidum in about half of the boxers studied,5,38 and electroencephalographic (EEG) abnormalities predominated by excessive slowing have been documented in approximately one-third to one-half of the boxers examined.38,40 Regional cerebral blood flow was found to be reduced in a series of professional fighters, but was normal in amateurs.37 Although neurological examination of boxers has generally revealed abnormalities less frequently than has CT or EEG recordings,5,38 Casson and colleagues5 reported that neuropsychological testing disclosed at least one “abnormal score” in all 18 cases they studied (including 13 participants who had retired from the sport at least 1 year before the study). Number of bouts, length of boxing career, and age at retirement have been positively related to the development of traumatic encephalopathy.5,35,38

Against a background of earlier neuropathological investigations demonstrating extensive histological and gross anatomical damage in boxers' brains,6,10,41 findings from the more recent neurobehavioral studies lend support to the concept of a continuum of brain injury which reflects the cumulative effects of repeated blows of subconcussive intensity and knockouts.5,35,38 The cumulative injury over a sufficient boxing experience may eventuate in progressive dementia, extrapyramidal, pyramidal, and cerebellar disorders, and personality changes in the most affected cases.6,9,16,41 In a study of 224 former boxers, Roberts35 found evidence of traumatic encephalopathy in 17% of the series.

The aforementioned studies implicate a high risk of neurobehavioral morbidity associated with extensive boxing experience, especially (but not exclusively) in professional bouts. However, these investigations were largely confined to ex-boxers who ranged in age up to 74 years,5,23,38 whereas relatively few boxers have been studied prospectively during the early stages of their amateur and professional careers.4,10,16 Although Casson, et al.,5 focused on boxers who were reportedly free of alcohol and drug abuse and other etiologies of neurological injury, some investigators21,32 have been less careful to exclude extraneous causes of neurobehavioral deficit. Variables such as sparring frequency, appropriateness of sparring partners and opponents, and intervals between bouts potentially contribute to morbidity.

Finally, utilizing published normative data to interpret neuropsychological test findings in boxers may introduce false-positive errors.5 It is questionable whether boxers are representative of the general population with respect to psychosocial variables. In a recent three-center study28 concerning outcome after mild head injury, significant geographic differences in neuropsychological performance were present across the samples of patients and their respective control groups despite apparent comparability in years of education. In the absence of suitably matched local control subjects, the investigators might have misinterpreted the data relative to the head-injured patients.

We have initiated a prospective longitudinal investigation of young amateur and professional boxers and a control group of amateur athletes matched on pertinent demographic variables. In this paper we report baseline and 6-month follow-up neuropsychological data on both groups and magnetic resonance imaging (MRI) findings in the boxers.

Clinical Material and Methods
Subject Selection

Managers of three boxing clubs in Houston, Texas, assisted in the recruitment of fighters whom they considered to have a long-term commitment to the sport. All of the boxers from two of the clubs were recruited. The manager of the third club recommended three of his boxers who exhibited a firm commitment to the sport (as reflected by regular training) and who were likely to continue fighting until the time of follow-up evaluation. In view of our focus on the effects of sustained boxing, exclusion of transient participants in the sport was deemed appropriate. We obtained a medical and social history from each boxer who consented to participate in the study. Exclusion criteria included a history of chronic alcohol or drug abuse, hospitalization for head injuries sustained outside the ring, or other neuropsychiatric disorders. Of the available pool of boxers, none declined our invitation to participate. Twelve professional boxers and three amateur boxers volunteered to participate in the study and satisfied the selection criteria. However, one of the professsionals dropped out before completing the examination and we subsequently deleted the data on one of the amateurs because he had only 3 months of organized boxing experience and no fights before entering the study. Although the boxers were not paid for participating in the study, the examiner provided them with feedback information concerning their performance. Table 1 summarizes the experience of the boxers in the ring.

TABLE 1

Amateur and professional experience of the 13 boxers in this series*

Boxer No.Yrs BoxingAmateur BoutsProfessional BoutsAmnesic EpisodesSparring Days/Mo
1 11150249
2‡ 81300512
 81155512
4 66949
5 91203510
6 61845315
7 8308518
8 5477315
9 54510315
10‡ 91001515
11 11959415
12 2142615
13‡ 7401115

Boxing experience before entry into the study. The 10 subjects who participated in follow-up testing continued to box with two to seven additional bouts since the baseline data were obtained. All boxers had winning records at the outset of the study and at follow-up examination, and none had suffered a knockout in a professional bout. Individual boxing records are not given so as to preserve the anonymity of the boxers. — = Amateur boxer at the time of this study.

The number of episodes of posttraumatic amnesia without loss of consciousness was determined by retrospective report.

These boxers failed to return for the 6-month neuropsychological examination.

This boxer declined to undergo magnetic resonance imaging but returned for 6-month neuropsychological testing.

Through newspaper and television advertisements, we recruited a comparison group of 13 young men from the Houston-Galveston area who regularly participated in an amateur sport other than boxing; these sports included amateur basketball (seven subjects), track (two subjects), baseball (one subject), football (one subject), karate (one subject), and weight lifting (one subject). We obtained medical and social histories to screen the control candidates according to the same criteria employed for the boxers. No control subject had a history of hospitalization for head injury or neuropsychiatric disorder. Apart from these exclusionary criteria, we selected the 13 men by individually matching them to one of the boxers according to age, years of education, parental socioeconomic background18 (Table 2), and race. The paternal occupations of the boxers and control subjects ranged predominantly from skilled manual jobs to unskilled workers. Occupations of the control subjects included five students, five cooks/food-service workers, one bookbinder, and two clerks. Each group consisted of nine Blacks, one Hispanic, and three Caucasians.

TABLE 2

Baseline demographic features and reading level of the boxers and control subjects*

FeatureBoxersControl Subjects
age (yrs)20.5 ± 2.120.0 ± 2.2
education (yrs)11.8 ± 1.511.7 ± 1.3
socioeconomic score38.2 ± 7.437.7 ± 9.7

Values are means ± standard deviations for 13 boxers and 13 control subjects.

Hollingshead index score of social class based on the father's occupation (or the mother's occupation if information on the father was unavailable).18 This social position score is the product of multiplying scaled values for occupation and education. Social class I (for instance, a judge) corresponds to a Hollingshead score range of 11–17, Class II = 18–27, Class III = 28–43, Class IV = 44–60, and Class V = 61–77 (for example, an unskilled employee with an eighth grade education).

Despite matching the boxers and control subjects according to the aforementioned variables, two differences emerged. Although we did not employ reading level as a matching variable, administration of the Wide Range Achievement Test20 disclosed a tendency for a lower oral reading level of single words by the boxers (mean grade level 7.2, standard deviation (SD) 1.9; mean percentile for age 28.1 years, SD 28.8) than by the control subjects (mean grade level 8.0, SD 3.0; mean percentile for age 46.6 years, SD 34.9). However, this unanticipated disparity between the groups was not significant (F(1,24) = (0.67: p > 0.05). An obvious difference between the two groups was their time spent in physical training. Although all control subjects participated regularly in an organized amateur sport, their daily regimen of total physical activity was less rigorous than the 3 to 5 miles of running, intensive gym exercise, and sparring which the boxers performed on a daily basis during training.

Attention/Information-Processing Speed
Paced Auditory Serial Addition Test

We administered a revision of the Paced Auditory Serial Addition Test (PASAT), which typically reveals slowing of information-processing speed during the initial hospitalization after a mild head injury.14,15 Although consecutive cases of mild head injury gradually improve to a normal level of performance on the PASAT within 1 to 3 months,14,15,28 recovery is delayed in patients with a moderate head injury and also in patients who have sustained multiple mild head injuries.14 The subject's task on the PASAT is to add each number to the preceding number (for instance, the correct responses for the series 1, 4, and 7 are 5 and 11); thus, the addition is serial but noncumulative. Four tape-recorded series of 50 single-digit numbers (requiring 49 additions) were presented, with an initial interstimulus interval of 2.4 seconds separating the successive numbers. The interstimulus interval was reduced by 0.4 seconds in each of three series (2.0 seconds, 1.6 seconds, and 1.2 seconds), which were separated by a pause of about 20 seconds. The PASAT was terminated when performance on any series fell below 50% correct. All of the study participants demonstrated adequate proficiency in arithmetic during unpaced practice trials to warrant administration of the PASAT.

Continuous Performance Test

Variations of the Continuous Performance Test (CPT) have been employed in numerous studies of attention deficit in patients with acquired brain damage and in psychophar-macological investigations.39 We employed an adaptive rate version of the CPT which was programmed on a microcomputer.33 The subject was instructed to press the return key as soon as a target (a silhouette of an airplane) appeared on the monitor, while withholding responses to distractors (five other designs). The program adjusted the rate of presentation of the stimuli according to the subject's performance;33 that is, it reduced the interstimulus interval by 10% (increasing the presentation rate) after a correct response to a target within the allotted time frame, while it increased the interstimulus interval by 20% (slowing the rate of presentation) following a false-positive error, such as responding to a distractor. After a series of practice trials, the subject completed four trial blocks (21 stimuli per block). The mean interstimulus interval for each trial block provided an index of selective attention. We were unable to administer the CPT at baseline to one of the control subjects. Consequently, his 6-month CPT data were also deleted from analysis.

Reaction Time

A microcomputer task to test reaction time, programmed by one of us (D.Z.), consisted of simple and complex conditions. In the simple condition, the same number (5) was presented in the center of the screen following a variable period (of 1 to 4 seconds) after a warning tone on each of 12 trials. The subject was instructed to lift his index finger from the home (letter “V”) key (which was depressed during the intertrial interval) and to press the “5” key as soon as possible when the number appeared on the screen. The intertrial interval was 4 seconds. Four practice trials were given to familiarize the subject with the task. In the complex condition, one of six numbers was presented on the monitor in a randomized sequence across 12 trials. The subject was instructed to lift his index finger from the home key and rapidly press the key corresponding to the number appearing on the monitor. The dependent measure in both the simple and complex conditions was time (measured in milliseconds) from the appearance of the number on the screen until the correct key was pressed. Technical problems precluded completion of this task at baseline by one of the three boxers who were subsequently lost to follow-up study. Recent research has demonstrated that relatively mild head injury results in a slowing of reaction time, which resolves within 6 months.30

Long-Term and Immediate Memory
Modified Selective Reminding

The modified selective-reminding procedure31 used in this investigation has been shown to characterize the initial impairment and recovery of long-term memory following mild head injury.28 Twenty names of common animals were read to the subject at the rate of one word every 2 seconds. Following presentation of the complete list, the subject was asked to recall the list in any order and he was encouraged to continue his recall after initial hesitation. After the subject indicated that he was unable to recall additional words, the examiner reminded him of only those items missed. The following trial began when the examiner asked the subject to again recall the entire list. This procedure was repeated over eight trials. Following termination of the procedure, a 60-minute de-layed-recall trial was given without prior warning. We analyzed the number of words consistently retrieved from long-term memory storage across trials and delayed recall.

Short-Term Visual Memory

We administered the Benton Visual Retention Test3 in which the examiner presented for 10 seconds each of 10 plates consisting of geometric designs. After the examiner withdrew the plate from view, the subject attempted to draw the design as accurately as possible. The performance measurements were the number of plates correctly reproduced (0 to 10) and the total number of errors. This standardized test of short-term visual memory was recently found to be highly sensitive to the effects of mild head injury.28

Digit Span Subtest of Immediate Memory

To contrast immediate recall with long-term memory, we administered the Digit Span subtest of the Wechsler Adult Intelligence Scale.42 Following the standard procedure for this subtest, we summed the maximum digit span forward and digit span backward of each subject and obtained a scale score and an age-correlated score.

Divergent Reasoning
Verbal Fluency

To assess verbal fluency, we used the controlled word-association test2 in which the subject retrieved as many words as possible beginning with a designated letter within 1 minute. This procedure was repeated for two other letters, and the total number of words retrieved (summed across the three letters) was analyzed. Benton2 has found that verbal fluency is disproportionately impaired in patients with left frontal lobe damage.

Design Fluency

The design fluency test is the visuospatial analog of the verbal fluency test. The subject was instructed to invent as many unique, novel designs as possible under time pressure. Jones-Gotman and Milner22 showed that patients undergoing right frontal lobectomy for relief of intractable epilepsy had impoverished productivity of designs, whereas their verbal fluency was relatively preserved. Following this procedure,22 we evaluated design fluency under a 5-minute free condition (that is, the only constraints being that the drawing could not be a scribble or a nameable figure) and a 4-minute fixed condition (that is, the designs had the additional constraint of being limited to four lines).

Visuomotor Coordination, Scanning, and Speed
Digit Symbol Test

The digit symbol test of visuomotor speed was recently shown to reflect slowing of visuomotor speed during the initial hospitalization for mild head injury.28 The examiner provided the subject with a reference key in which the numbers 1 through 9 were paired vertically with a distinctive nonverbal symbol.42 The test form consisted of a randomized sequence of the numbers that appeared above blank boxes. The subject's task was to write the missing symbols in the blank boxes corresponding to the number. The total number of correct symbols coded by the subject within the 90-second time limit was scored and transformed to a scale score and an age-corrected scale score.

Trail-Making Test

The trail-making test is a timed procedure1 that has been useful in characterizing the residual effects of severe closed head injury24 and the recovery curve after mild to moderate head injury.11 On Trail A the subject was requested to draw a line connecting 25 scattered circles in numerical order. Although the examiner pointed out errors and asked the subject to correct them, the primary dependent measure was time to completion. The complexity of the task was increased in Trail B, which required the subject to connect the scattered circles by drawing a line in alternating numerical and alphabetical sequences. As with Trail A, errors were corrected and the dependent measure was time to completion.

Purdue Pegboard Test

After a brief practice trial with the Purdue Pegboard, each hand was tested separately using a time limit of 30 seconds. Previous studies have established the sensitivity of this measure of visuomotor coordination and speed in instances of acquired brain damage.7

Interview

Each boxer and control subject was interviewed in order to assess complaints of postconcussional symptoms, cognitive problems, motor deficits, other somatic complaints, mood, and alterations in psychosocial functioning. The examiner also rated the presence of neurobehavioral disturbance.27

Magnetic Resonance Imaging

Magnetic resonance imaging was performed on a 0.35-Tesla unit within 2 months after the 6-month follow-up neuropsychological examination. Transaxial and coronal images were obtained using a 7-mm slice thickness with 3-mm intervals of nonimaged tissue between the slices and a spin-echo technique. The repetition time (TR) between successive spin-echo sequences was 2.0 or 3.0 seconds. The elapsed time (TE) between the 90° radiofrequency pulse and the reception of a spin-echo signal of the images ranged from 28 to 80 msec.

Results
Baseline Neurobehavioral Performance
Statistical Studies

Univariate and multivariate analysis of variance was used to compare the performance of the boxers with that of the control subjects. In view of the disparity in reading level between the boxers and the control athletes (which approached significance) and other individual differences that might predispose an individual to take up boxing, we employed a between-groups design rather than a matched-pairs analysis. Table 3 summarizes the baseline results, comparing the boxers and control subjects on each neuropsychological measure. To emphasize the presence of any differences in performance between the two groups, there is no correction in Table 3 for the number of statistical comparisons. Given the generally negative findings, a statistical correction would alter only the results for reaction time and the digit symbol (coding) test. Slight variation in degrees of freedom reflects the missing reaction time data for one boxer at baseline and the missing CPT data of one control subject on both examinations.

TABLE 3

Baseline neuropsychological test performance in boxers and control subjects*

Neuropsychological TestBoxersControl SubjectsGroup Effect
F(1.24)p Value   
attention/information-processing speed   
  PASAT optimal processing rate† 0.40 ± 0.15 0.41 ± 0.11 0.010.91
  reaction time (msec)   
    average reaction time 637.2 ± 82.2 738.5 ± 126.4 8.30.008
    simple reaction time 489.0 ± 71.8 609.1 ± 139.1 
    complex reaction time 785.4 ± 92.5 867.9 ± 113.7 
  Continuous Performance Test (CPT) interstimulus interval (msec)   
    average interstimulus interval 1006.9 ± 285.3 1039.8 ± 361.9 0.10.79
    Trial A 1102.9 ± 235.1 1128.3 ± 375.7 
    Trial B 883.2 ± 269.2 897.3 ± 306.1 
    Trial C 995.1 ± 318.1 991.9 ± 280.7 
    Trial D 1046.2 ± 318.1 1141.7 ± 484.9 
memory   
 verbal memory   
    average no. words retrieved 7.5 ± 3.4 10.5 ± 5.1 3.50.07
    60-minute recall 15.8 ± 2.8 16.2 ± 3.4 0.10.76
 visual memory   
    no. correct 6.8 ± 1.6 7.2 ± 2.0 0.20.66
    no. errors 4.2 ± 2.0 3.5 ± 2.5 0.60.44
 Digit Span Scale score 9.9 ± 2.5 11.0 ± 2.6 1.20.29
divergent thinking   
  verbal fluency (no. of words) 31.5 ± 10.7 38.2 ± 10.9 2.50.12
  design fluency   
    5-min free condition (no. of designs) 14.5 ± 5.0 18.4 ± 9.0 1.90.18
    4-min fixed condition (no. of designs) 11.6 ± 5.1 13.2 ± 4.9 0.60.44
visuomotor speed   
  Trail A time (sec) 24.0 ± 7.2 24.5 ± 7.6 0.040.85
  Trail B time (sec) 71.6 ± 24.6 78.2 ± 40.0 0.30.62
  Purdue Pegboard Test   
    preferred hand (no. of pegs) 14.8 ± 2.1 15.4 ± 1.8 0.50.49
    nonpreferred hand (no. of pegs) 13.5 ± 1.9 13.9 ± 1.8 0.30.60
  Digit Symbol Scale score 9.0 ± 1.4 10.8 ± 2.9 4.40.05
reading single words (percentile) 28.1 ± 28.8 46.6 ± 34.9 2.20.15

Values are means ± standard deviations for 13 boxers and 13 control subjects.

PASAT = Paced Auditory Serial Addition Test. Optimal processing rate was obtained for each subject by selecting the trial which yielded the highest processing rate (no. correct responses/total time of presentation).

Attention/Information-Processing Speed

The data from the PASAT were analyzed by computing the information-processing rate (that is, the number of correct responses/total time of presentation) for each of the four series. In view of the dropout of some subjects (both boxers and control subjects) in the series when the briefest interstimulus intervals (1.6 seconds and 1.2 seconds) were presented, we analyzed the PASAT optimal processing rate (that is, the fastest rate attained by each subject on any of the four trial blocks). The results disclosed no significant difference in information-processing rate between the boxers and the control subjects (Table 3).

The mean simple and complex reaction times of the boxers and control subjects are given in Table 3. We submitted the reaction times to a multivariate repeated-measures analysis of variance in which the group (boxers vs. controls) was a between-subjects variable and task complexity (simple vs. complex) was a within-subjects variable. As summarized in Table 3, this analysis confirmed shorter reaction times (that is, faster performance) in the boxers. The impression of faster reaction times under the simple condition was confirmed by a significant effect for task complexity (F(1,23) = 119.52; p < 0.001). Although there is a suggestion in Table 3 of a disproportionate increase in the response latencies of boxers under the complex condition, the interaction between the groups and task complexity was not significant (F(1,23) = 0.55; p < 0.47).

Sustained attention on the CPT was measured by the mean interstimulus interval on each of four trial blocks (Table 3). After an initial improvement in efficiency of responding (determined by a comparison of the inter-simulus interval on Trial Block A vs. Trial Block B), both groups exhibited a decline in performance as noted by the increased interstimulus intervals on Trial Blocks C and D as compared to B. This fluctuation in performance across trial blocks was highly significant (Wilks lambda = 0.234; Rao's F(3,21) = 22.92; p < 0.001). However, there was no difference in sustained attention between the boxers and the control individuals (Table 3), nor was the interaction between the groups and trials significant (Wilks lambda = 0.954; Rao's F(3,21) = 0.34; p = 0.80); in other words, there was no differential change in response efficiency between the boxers and the control subjects.

Long-Term and Immediate Memory

Table 3 summarizes the consistent long-term retrieval averaged across trials. Although a pattern of greater verbal learning and memory in the control subjects approached significance (Table 3), recall of the word list after a 60-minute delay was virtually identical in the two groups. In view of the ceiling effect associated with perfect recall by both groups on the eighth trial which eliminated any variance, we confined the analysis to seven trials. Learning of the word list was substantiated by a trials effect on recall (Wilks lambda = 0.067; Rao's F(6,19) = 44.63; p < 0.0001). Despite the apparent disparity between the learning curves of the boxers and control subjects across trials, the interaction of groups with trials was not significant (Wilks lambda = 0.769; Rao's F(6,19) = 0.96; p = 0.48).

Short-term visual memory for designs varied only slightly between the two groups (Table 3). Similarly, there was no significant difference between the boxers and control subjects with respect to the Digit Span Scale score (Table 3).

Divergent Reasoning

Control subjects tended to generate more words (verbal fluency) and more novel designs than the boxers. However, no group differences emerged from the analysis of variance.

Visuomotor Function

Table 3 summarizes the mean scores on the trail-making, digit symbol, and Purdue Pegboard tests. The directional trend of higher scores in the control subjects was unsubstantiated by the analysis of variance.

Interview Findings

The structured interview disclosed no marked behavioral disturbance as reflected by the examiner's ratings. Although somatic complaints and self-reporting of cognitive problems were generally minimal and did not differentiate the groups, there were two behavioral features noted in several of the boxers which were not apparent in the control athletes. Blunted affect (evidenced by reduced emotional tone, diminished intensity of feelings, and flatness of voice) was considered to be very mild or mild in four of the boxers and moderate in one. This was particularly evident during the formal examination as contrasted to their behavior during preliminary light conversation. Mild agitation, as reflected by restlessness, kicking, leg tapping, and picking, was observed in four of the boxers.

Summary of Baseline Findings

The baseline neurobehavioral scores tended to reflect higher performance in the control subjects (14 of the 23 measures in Table 3) as compared to the boxers. This trend was most evident in verbal learning. However, recall of the word list after a delay was essentially equal in the two groups, and no differences were present on other meassures of memory. Although the boxers had significantly faster reaction times than the control athletes, this finding would have been attenuated by correcting for the number of comparisons. No other differences between the groups were significant. An analysis of the effects of boxing experience on neurobehavioral functioning was equivocal because educational (r = 0.47, p < 0.05) and socioeconomic (r = 0.45, p < 0.06) levels were both positively related to number of bouts fought. Accordingly, several of the cognitive measures were positively related to the number of boxing matches.

Baseline Neurobehavioral Function Versus 6-Month Follow-up Function

Ten of the boxers were available for a 6-month follow-up examination. As indicated in Table 1, there was no consistent difference in boxing experience between the three boxers who failed to return for follow-up evaluation as compared to the other 10 boxers. We also found that compliance with follow-up examination was unrelated to the baseline neuropsychological test scores. Table 4 summarizes the serial neurobehavioral scores for the 10 boxers and of the corresponding control subjects who completed the 6-month follow-up examination.

TABLE 4

Baseline and follow-up neuropsychological test performance in serially studied boxers and control subjects*

Neuropsychological TestBaseline PerformanceFollow-Up PerformanceGroup EffectOccasion EffectGroup × Occasion
BoxersControl SubjectsBoxersControl SubjectsF(1,18)P ValueF(1,18)P ValueF(1,18)P Value 
attention/information-processing     
  PASAT optimal processing rate 0.41 ± 0.14 0.43 ± 0.11 0.46 ± 0.18 0.47 ± 0.14 0.10.794.70.050.040.85
  reaction time (msec)     
    average reaction time 642.6 ± 85.5 725.0 ± 118.1 645.6 ± 92.4 633.6 ± 99.8 0.90.358.00.019.10.007
    simple 500.4 ± 70.5 618.1 ± 157.0 513.3 ± 86.5 473.3 ± 111.3 
    complex 784.7 ± 100.5 831.8 ± 79.1 777.8 ± 98.3 793.9 ± 88.3 
  Continuous Performance Test (CPT) interstimulus interval (msec)     
    average interstimulus interval 987.1 ± 303.7 1078.4 ± 404.8 934.2 ± 327.0 877.5 ± 199.6 0.010.894.20.061.50.24
    Trial A 1069.5 ± 247.5 1149.1 ± 428.1 987.4 ± 228.0 988.1 ± 231.8 
    Trial B 863.9 ± 272.2 937.8 ± 337.3 881.0 ± 408.6 819.2 ± 230.6 
    Trial C 998.7 ± 349.7 1022.1 ± 305.7 953.4 ± 337.3 820.4 ± 149.4 
    Trial D 1016.2 ± 345.2 1204.4 ± 548.0 915.1 ± 334.2 882.3 ± 186.5 
memory     
  verbal memory     
    average no. words retrieved 7.9 ± 3.7 11.7 ± 3.8 11.0 ± 3.7 12.7 ± 5.2 2.80.119.30.0072.70.12
    60-minute recall 16.0 ± 3.1 16.8 ± 3.2 18.2 ± 1.8 18.6 ± 2.1 0.40.539.20.0070.10.77
  visual memory     
    no. correct 6.8 ± 1.6 7.4 ± 2.2 7.2 ± 1.4 8.3 ± 1.5 1.70.212.90.110.40.52
    no. errors 4.2 ± 2.0 2.8 ± 2.5 3.3 ± 1.9 2.5 ± 2.8 1.50.241.30.260.30.57
  Digit Span Scale score 10.1 ± 2.6 11.3 ± 2.8 10.5 ± 2.4 10.9 ± 2.3 0.60.4501.00.90.34
divergent thinking     
  verbal fluency (no. of words) 31.5 ± 10.7 37.0 ± 11.4 24.9 ± 19.6 37.3 ± 12.8 2.90.101.50.231.40.25
  design fluency     
    free cond. (no. of designs) 14.4 ± 4.8 19.2 ± 8.9 16.7 ± 11.7 23.6 ± 11.4 2.160.164.30.050.420.53
    fixed cond. (no. of designs) 10.4 ± 5.2 13.8 ± 3.6 11.2 ± 6.8 16.3 ± 4.9 3.680.074.40.051.170.29
visuomotor speed     
  Trail A (sec) 25.0 ± 7.7 24.1 ± 8.5 22.1 ± 5.1 26.3 ± 6.8 0.40.560.050.832.60.12
  Trail B (sec) 71.9 ± 28.0 66.5 ± 38.2 61.1 ± 19.7 58.2 ± 33.4 0.100.754.40.050.070.79
  Purdue Pegboard Test     
    preferred hand (no. of pegs) 14.8 ± 2.3 15.5 ± 1.8 15.0 ± 1.8 16.2 ± 1.5 2.00.180.80.390.20.63
    nonpreferred hand (no. of pegs) 13.4 ± 1.6 13.5 ± 1.8 13.7 ± 1.5 14.8 ± 1.3 1.10.324.60.051.80.20
  Digit Symbol Scale score 8.8 ± 1.1 10.8 ± 2.9 11.0 ± 1.5 10.9 ± 2.5 1.70.213.40.082.90.11
reading single words (percentile) 30.5 ± 31.9 52.3 ± 34.9 31.4 ± 33.1 57.0 ± 36.7 2.50.130.70.410.30.57

Baseline and follow-up performance values are means ± standard deviations for 10 boxers and 10 control subjects evaluated at each examination time.

PASAT = Paced Auditory Serial Addition Test. Optimal processing rate was obtained for each subject by selecting the trial which yielded the highest processing rate (no. correct responses/total time of presentation).

Attention/Information-Processing Speed.

Table 4 depicts a pattern of generally improved performance in attention/information-processing speed on the follow-up examination. Information-processing rate, as meassured by the PASAT, significantly improved as compared to the baseline scores. However, there was no significant difference in information-processing rate between the boxers and control subjects, nor was there any differential change in performance between the two groups; that is, there was no interaction of groups and occasions (Table 4).

The mean interstimulus intervals on the CPT showed no impressive group differences, whereas the trend toward more efficient performance on follow-up examination as compared to baseline measurements approached significance (Table 4). There was no differential change from baseline findings to results at 6 months in sustained attention in the boxers as compared to the control group (that is, there was a nonsignificant interaction of groups and occasions). Consistent with the findings obtained in the mean baseline results of all 13 individuals in each of the two groups, serial testing confirmed a highly significant fluctuation in performance across trials; that is, an initial improvement in response efficiency followed by a decrement in responding (Wilks lambda = 0.2272; Rao's F(3,15) = 17.01; p = 0.001). However, there was no interaction of groups with trials (Wilks lambda = 0.7534; Rao's F(3,15) = 1.64; p = 0.22); in other words, there was no differential fluctuation in performance across trials between the groups. Variation in response efficiency across trials also did not differ between the baseline and the follow-up examinations (Wilks lambda = 0.7201; Rao's F(3,15) = 1.94; p = 0.17), nor was the three-way interaction of groups, occasions, and trials significant (Wilks lambda = 0.9448; Rao's F(3,15) = 0.29; p = 0.83).

Improved performance on the follow-up examination was also confirmed on the reaction-time test (see Table 4). In contrast to the shorter reaction times displayed by the boxers as compared to the control subjects on the baseline examination, Fig. 1 indicates that the disparity in response latency between the two groups was considerably narrowed at the time of their follow-up examination. The impression of greater improvement in reaction time on the follow-up examination by the control subjects as compared to the boxers was substantiated by the significant interaction of groups and occasions (Table 4). As is evident in Fig. 1, reaction times under the complex condition were slower than under the simple condition (F(1,18) = 205.76; p < 0.001). However, there was no interaction of the groups and task complexity (F(1,18) = 0.04; p < 0.85).

Fig. 1.
Fig. 1.

Graph showing the mean simple and complex reaction times plotted for the boxers and control subjects who were examined at both the baseline and the 6-month follow-up examinations. For a description of simple and complex reaction times see text.

Long-Term and Immediate Memory

Consistent long-term retrieval on the verbal memory test is plotted against trials for both the baseline and follow-up examinations in Fig. 2. In contrast to the trend toward greater acquisition of the word list by the control subjects on the baseline test, the follow-up data show a narrowing of the gap between the groups. Multivariate analysis of the verbal memory data disclosed no significant difference between the groups, whereas an improvement in performance on follow-up evaluation was confirmed (Table 4). The trend of greater improvement from baseline to 6 months by the boxers as compared to controls fell short of significance (that is, there was no interaction of groups and occasions). As reflected in Fig. 2, consistent retrieval of words increased over trials (Wilks lambda = 0.0226; Rao's F(6,13) = 93.80; p < 0.001). There was no interaction of groups and trials (Wilks lambda = 0.8174; Rao's F(6,13) = 48; p = 0.81). Figure 2 reveals that the word-acquisition curve of the boxers was steeper on follow-up examination than their learning curve at baseline, whereas the control athletes exhibited a relatively minor change in their rate of learning. Consistent with this impression, there was a significant three-way interaction of groups, occasions, and trials (Wilks lambda = 0.3028; Rao's F(6,13) = 4.99; p = 0.007).

Fig. 2.
Fig. 2.

Graph showing consistent long-term retrieval of words plotted across trials on the modified selective-reminding test given at baseline and at the follow-up examination. Recall after a 60-minute delay is also shown. For a description of the long-term word-retrieval procedure see text.

Figure 2 and Table 4 show a significant trend toward improvement in delayed recall of the word list on follow-up examination as compared to baseline. At the 6-month examination, both groups correctly recalled an average of about 18 of the 20 words after 60 minutes of intervening testing (Table 4). Table 4 summarizes baseline and follow-up findings on the Benton Visual Retention Test and the Digit Span Scale score. There were no group differences, no significant change in performance between the baseline and follow-up examinations, and no interactions.

Divergent Reasoning

Table 4 reveals no significant group differences on either the verbal or design fluency tests. Although there was no significant change from baseline to follow-up evaluation nor any interactions, verbal fluency tended to be more variable in the boxers on the second test than on the first examination.

Visuomotor Function

The visuomotor test data are summarized in Table 4. There were no significant differences in performance between the boxers and the control subjects on any of these measures, nor was there any differential improvement (that is, an interaction of groups and occasions) in scores between the baseline and follow-up examinations. Improved performance on the 6-month examination was confined to Trail B and pegboard assembly using the nonpreferred hand.

Interview Findings

The follow-up interview observations were generally consistent with the behavior recorded during the baseline interviews. There were no impressive changes noted as compared to the baseline examination.

Magnetic Resonance Imaging

With the exception of Boxer 3 (see Table 1), all of the boxers who returned for a 6-month neuropsychological examination consented to undergo MRI. The images of all nine boxers were interpreted as normal by two neuroradiologists (A.G. and S.H.). Nine transaxial slices are shown for two of the boxers in Figs. 3 and 4. Neither neuroradiologist interpreted any of these magnetic resonance images as demonstrating evidence of increased sulcal prominence, ventricular enlargement suggestive of cerebral atrophy, or cavum septum pellucidum.

Fig. 3.
Fig. 3.

Magnetic resonance images in the transaxial plane for Boxer 11 (see Table 1) obtained 3 months after his 6-month follow-up examination.

Fig. 4.
Fig. 4.

Magnetic resonance images in the transaxial plane for Boxer 12 (see Table 1) obtained 1 month after his 6-month follow-up examination.

Discussion

Our study differs from previous research5,23,38 concerning the neurobehavioral sequelae of boxers, in that we recruited individuals under 30 years of age who were entering the early stages of a professional career or who were still amateurs. Moreover, all participants were competitive in the ring at the time of the study as opposed to previous reports on former boxers who were at risk for repeated brain injury during the declining stages of their career. Apart from differences in the selection of boxers, we individually matched a control subject engaged in amateur athletics to each boxer based on demographic variables and we retested 10 of the subjects in each group after 6 months. Admittedly, recruitment of professional athletes for the control group would have provided a more appropriate comparison group for the 11 boxers who were no longer amateurs at the time of this study. Although it was not feasible for us to blind the examiners to the group membership of the boxers and control athletes, we included two tasks (reaction-time test and CPT) which were administered and scored by a microcomputer, and a tape-recorded task (PASAT) which at least mitigated the potential effects of experimenter bias. Although we recruited boxers from three clubs who had demonstrated to their managers a strong commitment to remaining in the sport, it was difficult to evaluate the degree to which these men were representative of individuals engaged in an early stage of a boxing career. Finally, caution is advised in drawing inferences from our data pending replication and follow-up study over a longer interval.

The neuropsychological test data show a trend toward deficient reading and verbal learning in the boxers. Although the boxers more closely approximated the verbal learning scores of the control individuals at the 6-month follow-up examination, their reading scores still tended to be lower. It is conceivable that a subtle impairment in processing verbal material resulted from boxing in these young men. Other plausible interpretations include an association between self-selection for boxing and a subtle learning disability or at least subjective difficulty in reading. However, it is important to recognize the wide variation in reading ability among the boxers studied. In fact, two of the boxers were undergraduate college students maintaining a B average at the time of the study. The essentially identical delayed recall in both groups and the negative findings on the Benton Visual Retention Test provide no support for a pervasive memory disturbance in the boxers.

At the 6-month follow-up examination our finding of faster reaction time in the 13 boxers at the baseline examination was diminished and no longer significant. Although there were directional trends on other neuropsychological tests, none were significant at the time of either the baseline or the follow-up assessment.

Boxers and control subjects tended to improve their performance or at least remain stable on the follow-up examination. In view of the continued sparring and additional bouts contested by the boxers during the interim, differential change (for example, decline in the boxers as opposed to gains in the control subjects) would have been anticipated. Notwithstanding the lack of evidence for deterioration over a 6-month period in the boxers on tests of cognition, memory, or motor speed, we recognize that longitudinal investigation over an extended period might disclose signs of delayed neurological disorder. Previous neuropathological investigations have alluded to or implicated the delayed appearance of neurological symptoms following intensive boxing experience during adolescence and young adulthood.6,9,32 Moreover, retrospective epidemiological studies of clinically diagnosed Alzheimer's disease have identified head injury as the life event most closely associated with development of this form of dementia.12,17,36 In contrast to earlier studies, we were unable to confirm an inverse relationship between the number of bouts and the level of performance on neuropsychological tests. However, the confounding of education and parental socioeconomic level with boxing experience and the relatively few professional bouts fought may account for these negative findings.

Whether the mild affective blunting observed during the examination reflects a learned strategy of coping with stressful (or painful) conditions, the effects of brain injury, or individual differences which predispose to a boxing career awaits further investigation. Although the mild motor restlessness exhibited by the boxers could conceivably reflect “latent” extrapyramidal injury,9 this interpretation is unsubstantiated by their psychomotor data.

Magnetic resonance imaging produced normal findings in all nine boxers studied with this technique. Moreover, these individuals returned for follow-up examination and showed no evidence of cognitive deterioration. Consistent with our MRI data, Smith and Haughton (unpublished data, 1986) recently found no evidence of cerebral atrophy on MRI studies in nine boxers. However, their 10th subject, a 65-year-old man who retired after 101 professional bouts, was demented and had large ventricles. Recent studies have demonstrated the capability of MRI to detect cerebral lesions that are missed completely or appear smaller on CT scanning in patients who have sustained mild closed head injury.13,26

Notwithstanding the constraints imposed on extrapolating from this small serially studied group, our findings raise the possibility that young boxers may escape disabling brain injury provided that their exposure in the ring is limited both in frequency and total duration. Substantially reducing the number of rounds in professional bouts and strictly enforcing medical regulations in all states could mitigate or prevent development of dementia and other neurological symptoms in boxers. Periodic neurobehavioral and neurological examinations using a uniform protocol supplemented by MRI and EEG recordings could potentially detect impending neurological deterioration due to cumulative brain injury.41

Acknowledgments

We are indebted to Lori Bertolino for assistance in testing the subjects and data analysis, to Beverly Parman and Liz Zindler for manuscript preparation, and to Dr. Arthur L. Benton and Dr. Barry Jordan for their critique of the paper. We are also grateful to the boxers who volunteered to participate in this investigation.

References

This study was supported by a grant from the J.S. Aber-crombie Foundation and National Institutes of Health Grant NS-21889.

These data were presented in part at the World Medical Congress on Olympic Boxing in Reno, Nevada, on May 10, 1986.

Article Information

Address reprint requests to: Harvey S. Levin, Ph.D., Division of Neurosurgery D-73, The University of Texas Medical Branch, Galveston, Texas 77550.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Graph showing the mean simple and complex reaction times plotted for the boxers and control subjects who were examined at both the baseline and the 6-month follow-up examinations. For a description of simple and complex reaction times see text.

  • View in gallery

    Graph showing consistent long-term retrieval of words plotted across trials on the modified selective-reminding test given at baseline and at the follow-up examination. Recall after a 60-minute delay is also shown. For a description of the long-term word-retrieval procedure see text.

  • View in gallery

    Magnetic resonance images in the transaxial plane for Boxer 11 (see Table 1) obtained 3 months after his 6-month follow-up examination.

  • View in gallery

    Magnetic resonance images in the transaxial plane for Boxer 12 (see Table 1) obtained 1 month after his 6-month follow-up examination.

References

1.

Army Individual Test Battery. Manual of directions and scoring. Washington, DC: War Department, Adjutant General's Office1944Army Individual Test Battery. Manual of directions and scoring.

2.

Benton AL: Differential behavioral effects in frontal lobe disease. Neuropsychologia 6:53601968Benton AL: Differential behavioral effects in frontal lobe disease. Neuropsychologia 6:

3.

Benton AL: The Revised Visual Retention Test: Clinical and Experimental Applicationsed 4. New York: The Psychological Corporation1974Benton AL:

4.

Blonstein JLClarke E: Further observations on the medical aspects of amateur boxing. Br Med J 1:3623641957Br Med J 1:

5.

Casson IRSiegel OSham Ret al: Brain damage in modern boxers. JAMA 251:266326671984JAMA 251:

6.

Corsellis JANBruton CJFreeman-Browne D: The aftermath of boxing. Psychol Med 3:2703031973Psychol Med 3:

7.

Costa LDVaughan HG JrLevita Eet al: Purdue Pegboard as a predictor of the presence of laterality of cerebral lesions. J Consult Psychol 27:1331371963J Consult Psychol 27:

8.

Council on Scientific Affairs: Brain injury in boxing. JAMA 249:2542571983Council on Scientific Affairs: Brain injury in boxing. JAMA 249:

9.

Critchley M: Medical aspects of boxing, particularly from a neurological standpoint. Br Med J 1:3573621957Br Med J 1:

10.

Cruikshank JKHiggens CSGray JR: Two cases of acute intracranial haemorrhage in young amateur boxers. Lancet 1:6266271980Lancet 1:

11.

Dikmen SReitan RMTemkin NR: Neuropsychological recovery in head injury. Arch Neurol 40:3333381983Arch Neurol 40:

12.

French LRSchuman LMMortimer JAet al: A case-control study of dementia of the Alzheimer type. Am J Epidemiol 121:4144211985Am J Epidemiol 121:

13.

Gandy SESnow RMZimmerman RDet al: Cranial nuclear magnetic resonance imaging in head trauma. Ann Neurol 16:2542571984Ann Neurol 16:

14.

Gronwall DWrightson P: Cumulative effect of concussion. Lancet 2:9959971975Lancet 2:

15.

Gronwall DWrightson P: Delayed recovery of intellectual function after minor head injury. Lancet 2:6056091974Lancet 2:

16.

Harvey PKPDavis JN: Traumatic encephalopathy in a young boxer. Lancet 2:9289291974Lancet 2:

17.

Heyman AWilkinson WEStafford JAet al: Alzheimer's disease: a study of epidemiological aspects. Ann Neurol 15:3353411984Ann Neurol 15:

18.

Hollingshead AB: Two Factor Index of Social Position. New Haven: AB Hollingshead1957Hollingshead AB: Two Factor Index of Social Position.

19.

Ioannou S: Should boxing be banned? Can Med Assoc J 131:101984 (Letter)Ioannou S: Should boxing be banned? Can Med Assoc J 131:

20.

Jastak JFJastak S: The Wide Range Achievement Test: Manual of Instructions (1978 Revised Edition). Wilmington, Del: Jastak Associates, Inc1978The Wide Range Achievement Test: Manual of Instructions (1978 Revised Edition).

21.

Johnson J: Organic psychosyndromes due to boxing. Br J Psychiatry 115:45531969Johnson J: Organic psychosyndromes due to boxing. Br J Psychiatry 115:

22.

Jones-Gotman MMilner B: Design fluency: the invention of nonsense drawings after focal cortical lesions. Neuropsychologia 15:6536741977Neuropsychologia 15:

23.

Kaste MKuurne TVilkki Jet al: Is chronic brain damage in boxing a hazard of the past? Lancet 2:118611881982Lancet 2:

24.

Klove HCleeland CS: The relationship of neuropsychological impairment to other indices of severity of head injury. Scand J Rehab Med 4:55601972Scand J Rehab Med 4:

25.

Lampert PWHardman JM: Morphological changes in brains of boxers. JAMA 251:267626791984JAMA 251:

26.

Levin HSAmparo EEisenberg HMet al: Magnetic resonance imaging and computerized tomography in relation to the neurobehavioral sequelae of mild and moderate head injuries. J Neurosurg 66:7067131987J Neurosurg 66:

27.

Levin HSHigh WMGoethe KEet al: The neurobehavioral rating scale: assessment of the behavioral sequelae of head injury by the clinician. J Neurol Neurosurg Psychiatry 50:1831931987J Neurol Neurosurg Psychiatry 50:

28.

Levin HSMattis SRuff RMet al: Neurobehavioral outcome following minor head injury: a three-center study. J Neurosurg 66:2342431987J Neurosurg 66:

29.

Lundberg GD: Boxing should be banned in civilized countries. JAMA 249:2501983Lundberg GD: Boxing should be banned in civilized countries. JAMA 249:

30.

MacFlynn GMontgomery EAFenton GWet al: Measurement of reaction time following minor head injury. J Neurol Neurosurg Psychiatry 47:132613311984J Neurol Neurosurg Psychiatry 47:

31.

Mattis SKovner R: Different patterns of mnemonic deficits in two organic amnestic syndromes. Brain Lang 6:1791911978Brain Lang 6:

32.

Mawdsley CFerguson FR: Neurological disease in boxers. Lancet 2:7958011963Lancet 2:

33.

Murphy-Berman VWright GRosell J: Measuring children's attention span: a microcomputer assessment technique. J Educ Res 80:23281986J Educ Res 80:

34.

Patterson RH Jr: On boxing and liberty. JAMA 255:248124821986Patterson RH Jr: On boxing and liberty. JAMA 255:

35.

Roberts AH: Brain Damage in Boxers. A Study of the Prevalence of Traumatic Encephalopathy Among Ex-Professional Boxers. London: Pitman Medical and Scientific1969Roberts AH: Brain Damage in Boxers. A Study of the Prevalence of Traumatic Encephalopathy Among Ex-Professional Boxers.

36.

Rocca WAAmaducci LASchoenberg BS: Epidemiology of clinically diagnosed Alzheimer's disease. Ann Neurol 19:4154241986Ann Neurol 19:

37.

Rodriguez GFerrillo FMontano Vet al: Regional cerebral blood flow in boxers. Lancet 2:8581983 (Letter)Lancet 2:

38.

Ross RJCole MThompson JSet al: Boxers — computed tomography, EEG and neurological evaluation. JAMA 249:2112131983JAMA 249:

39.

Rosvold HEMirsky AFSarason Let al: A continuous performance test of brain damage. J Consult Clin Psychol 20:3433501956J Consult Clin Psychol 20:

40.

Sironi VAScotti GRavagnati Let al: CT-scan and EEG findings in professional pugilists: early detection of cerebral atrophy in young boxers. J Neurosurg Sci 26:1651681982J Neurosurg Sci 26:

41.

Uterharnscheidt F: Boxing injuriesSchneider RCKennedy JCPlant ML (eds): Sports Injuries. Baltimore: Williams & Wilkins1985Sports Injuries.

42.

Wechsler D: Wechsler Adult Intelligence Scale Manual. New York: The Psychological Corporation1955Wechsler D: Wechsler Adult Intelligence Scale Manual.

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 2 2 2
Full Text Views 55 55 25
PDF Downloads 67 67 39
EPUB Downloads 0 0 0

PubMed

Google Scholar