in mild traumatic brain injury

BRAIN INJURY ,

2002,

VOL.

16,

NO.

3, 185± 195

Remediation of `working attention’ in mild traumatic brain injury KEITH D. CICERONE JFK-Johnson Rehabilitation Institute, Edison, NJ, USA (Received 19 March 2001; accepted 4 October 2001 ) Several studies have reported beneficial effects of treatments for attentional deficits following traumatic brain injury. Improvements in speed of processing appear to be less robust than improvements on nonspeeded tasks, while several studies suggest greater benefits of training more complex forms of attention. The present study presents preliminary results concerning the effectiveness of an intervention for attentional deficits after mild traumatic brain injury. The treatment was based upon the conceptualization of deficits and interventions as a function of the central executive component of working memory, or `working attention’. A prospective, case-comparison design was employed comparing four treatment participants with an untreated comparison sample. Treatment tasks were derived from experimental procedures which have been demonstrated to elicit working memory demands, consisting of `n-back’, random generation, and dual-task procedures. The intervention emphasized the conscious and deliberate use of strategies to effectively allocate attentional resources and manage the rate of information during task performance. Treatment participants were more likely to exhibit clinically significant improvement on measures of attention and reduction of self-reported attentional difficulties in their daily functioning. Further analysis suggested that the principal effect of the intervention was on working memory, i.e. the ability to temporarily maintain and manipulate information during task performance, with no direct effect on processing speed. The results are consistent with a strategy training model of remediation, in which the benefits of treatment are due to participants’ improved ability to compensate for residual deficits and adopt strategies for the more effective allocation of their remaining attentional resources.

Introduction Several studies have reported beneficial effects of treatments for attentional deficits following traumatic brain injury (TBI), although none of these studies allows for an evalution of the specific effectiveness of attention training for mild traumatic brain injury (MTBI). Mateer et al. [1] described five cases with mild concussive or whiplash injuries who received treatment 1± 5 years after their injury. The postacute rehabilitation programme included individual and group cognitive remediation for attention and memory deficits over a period of 8 months. Following treatment, these five subjects made clinically significant gains on 60% of the neuropsychological measures of attention, memory and general intellectual functioning administered. Gray et al. [2] treated 31 TBI and stroke subjects with attentional dysfunction (based on subjective report and impaired performance on a screening measure of Correspondence to: Keith D. Cicerone, PhD, JFK-Johnson Rehabilitation Institute, 2048 Oak Tree Road, Edison, NJ 08820, USA. e-mail: [email protected] Brain Injury ISSN 0269± 9052 print/ISSN 1362± 301X online # 2002 Taylor & Francis Ltd http://www.tandf.co.uk/journals DOI: 10.1080/02699050110103959

186

K. D. Cicerone

attention) randomly assigned to attention retraining or recreational computing. The attentional training was based on the treatment of control processes, e.g. working memory or dividing attention, including simultaneous dual tasks and allocation of attention in complex situations. The attention treatment produced improvement relative to the control group on two measures of attention. Effectiveness was not related to the severity of attentional impairment prior to treatment. At 6-month follow-up, the treatment group demonstrated continued improvement and superior performance compared with the control group on tests involving auditory-verbal working memory. Niemann et al. [3] evaluated an attention training programme, which included divided attention tasks intended to improve the allocation of attentional resources, for patients with TBI. Following treatment, the attention training group improved significantly more than an alternative treatment group on attention measures administered throughout the treatment period, although the effects did not generalize to the second set of neuropsychological measures. Both treatment groups exhibited some improvement on all measures, but the effects on specific measures were weak. There have been several attempts to establish the differential effectiveness of training for specific components of attention. Improvements in speed of processing appear to be less robust than improvements on non-speeded tasks, while several studies suggest greater benefits of training more complex forms of attention [2, 4, 5]. Recent neuropsychological evidence indicates that the attentional disturbance MTBI is due, in addition to slowed processing speed, to an impairment of working memory [6, 7]. This impairment may reflect dysfunction of the anterior attentional system involved in the `central executive’ component of working memory [8, 9]. According to the model proposed by Baddeley and Hitch [10], working memory is comprised of two slave systems, the visuo-spatial sketchpad and the phonological loop, capable of holding and manipulating visuospatial or speech-based information, respectively. The central executive component plays a supervisory role over these modality-specific systems and is principally concerned with the integration and ongoing control of information. It is interesting to note that Baddeley [11] has suggested that the central executive component of working memory might better be considered a process of `working attention’. The present paper describes and presents preliminary results concerning the effectiveness of an intervention for attentional deficits after MTBI, based upon the conceptualization of deficits and interventions as a function of the central executive component of working memory, or `working attention’. The design of the intervention was based on several clinical assumptions. The subjective complaints and functional difficulties expressed by patients with MTBI were considered to reflect problems with the temporary maintenance and manipulation of mental representations of information. These difficulties were felt to be more pronounced with demands to attend to rapidly presented information and to attend to multiple sources of information, conditions which required more efficient allocation of attentional resources. As a result of this impairment, these patients’ experienced their cognitive functioning as more effortful and less automatic. Therefore, exposure and practice on tasks with demonstrated working memory demands would provide patients with the opportunity to use strategies to improve their regulation and efficient allocation of attentional resources.

Remediation of `working attention’

187

Method Participants A prospective, case-comparison design was employed with a convenience sample of patients referred to a post-acute brain injury rehabilitation programme for evaluation and treatment secondary to diagnosis of MTBI. Participants were at least 3 months post-injury, to minimize the effects of neurologic recovery. Patients with evidence of incomplete effort on neuropsychological testing, those with significant emotional or psychiatric disturbance, as well as patients whose symptomatic complaints and disturbance in social and occupational functioning appeared grossly disproportionate to their injuries, were considered inappropriate for either the treatment or comparison sample. Participants were selected based on evidence for impaired performance (t-score < 40) on at least two of six pre-treatment attention measures, and subjective report of attentional difficulties in their daily functioning (total score greater than 1 SD above the normal mean on a self-report measure, the Attention Rating and Monitoring Scale). The results of treatment for four participants were compared with four participants who were administered the appropriate pre- and post-treatment measures within a comparable time frame, and were referred for treatment but unable to participate. Two of the comparison participants were unable to enter treatment due to their geographic distance from the treatment facility, one participant’s treatment was postponed due to funding, and one participant’s treatment was postponed while awaiting recuperation from orthopaedic surgery. The treatment and comparison participants were matched, as closely as possible, on gender (three females, one male), age (mean = 31.0 vs. 34.75 years), education (15.25 vs. 15.25 years), months post-injury (8.25 vs. 7.0), and pre± post testing interval (22.25 vs. 24.25 weeks). Procedures Treatment tasks were derived from experimental procedures which have been demonstrated to elicit working memory demands. The tasks were administered in a hierarchical sequence. This hierarchy was not based on any theoretical rationale regarding the structure of attention, but was rather based upon a rational sequence of task difficulty with an attempt to establish competence at a lower level of task demands before introducing more complex task demands or additional task components. Three general levels of task complexity were utilized. N-back procedures The n-back task has previously been used to examine the neural basis of the central executive component of working memory [8]. The general n-back procedure consists of the presentation of a sequence of stimuli with the requirements for the participant to continuously report the stimulus occurring n number of stimuli previously. For example, in the 1-back condition a set of digits are presented sequentially in random order and the participant is instructed to report the digit which occurred one prior to the current digit. The 2-back condition also consists of a randomly ordered sequence of digits, however the participant is now instructed to report the digit which occurred once removed from (or two prior to) the digit currently presented (figure 1). While this paradigm can be presented using various

188

K. D. Cicerone

Stimuli:

Response:

--

--

“10”

“2”

“6”

Figure 1. Illustration of `2-back’ task. Participant was required to name the stimulus card that occurred two cards prior to the card currently exposed. In the typical task used in treatment, only the most current stimulus card was exposed, and the participant had to maintain a running mental representation of previous stimuli and recall the appropriate response from memory.

modalities and protocols, the current treatment procedure was implemented with the use of common playing cards as a means of presenting the stimuli. This was done for two reasons. First, it was felt that the use of playing cards would provide a familiar, non-threatening context for the treatment participants. Secondly, the manual presentation of stimuli allowed the therapist to quickly and easily modify conditions of the task, such as changing from self-paced to externally paced presentation or reviewing previously viewed stimuli. Three variants of the basic n-back task were presented. (a) Thirty-eight common playing cards from one suit were presented in random sequence, with only the current card showing, and participants required to report either the 1-back, 2-back, or 3-back card value in the sequence. The presentation of cards could be either experimenter-paced, or self-paced. (b) Thirty-eight common playing cards from two suits of different colours (i.e. red or black) were presented in random order as in the above condition, and participants were again required to report the n-back card in the sequence. In addition, working memory `load’ was increased by requiring participants to name either the colour of the card currently showing, or the colour of the card being named, before reporting the card value. (c) Thirty-eight playing cards from all four suits were presented, and participants were required to sort the cards into four piles by suit while simultaneously reporting the n-back card value. Thus, there were four card values showing at any time, and the card value which was to be named could either be exposed or covered by another card at different sequences in the sort. Preliminary use of this variant of the task indicated that the additional


189

exposed cards served as a source of potential interference. This version of the task could again be experimenter- or self-paced. N-back with intermittent verbal generation The primary n-back tasks were identical to those described above. In addition, participants were required to make a self-generated response on each trial, prior to naming the relevant card in the n-back task. This condition was introduced to further increase demands upon working memory and, specifically, to simulate the functional requirement of internally generating a response while maintaining a mental representation which would facilitate resumption of the (external) n-back task. Two variants of this condition were employed, based upon a random generation procedure which has previously been employed to assess working memory in persons with TBI [12]. In the first condition, participants were required to generate an example from two or more semantic categories, without repeating examples from the same category on successive trials. In the second condition, participants were required to generate a random letter triad (e.g. AKU) without using examples which represented words, natural letter sequences, or acronyms. N-back with continuous secondary task In the third condition, participants were engaged in an ongoing secondary task while maintaining performance on the primary n-back task and actively allocating attentional resources between the various task demands. This condition was introduced to simulate the occurrence of interrupting an ongoing activity in order to respond to an additional task, while maintaining a mental representation of the primary task which would allow one to return to the original activity. The secondary tasks were individually tailored in an effort to replicate each participants real-life demands. For example, in the case of one participant who was required to participate in conference phone-calls and negotiate the needs of several clients as part of her work responsibilities, the secondary task consisted of shadowing audiotaped lectures and conversations; another participant whose work required him to look up serial numbers from an inventory and enter these numbers into another database for requisitions, was required to perform an ongoing, written clerical task to determine if two numbers and names were identical. The specific timing and nature of interventional tasks was varied according to individual clinical needs. Although the introduction of task conditions was generally sequential, participants were at times requested to return to an earlier task. The rationale for this was twofold. First, it was felt important to reinforce a level of proficiency which had already been demonstrated. Secondly, the need to perform tasks with varied levels of complexity was employed to address participant’s ability to allocate their attentional resources in accordance with different degrees of task demands, as well as to self-monitor and self-regulate their subjective effort and likelihood of errors. All of the treatments were conducted individually, 1 hour each week, for a period ranging from 11± 27 weeks for the four treatment participants. Within each 1-hour session, 20± 30 minutes were devoted to feedback and discussion of the participant’s performance, identification of task variables which influenced performance, development of strategies for effective task performance, management of secondary emotional responses during task performance (e.g. frustration), especially insofar as these reactions interfered with cognitive performance, identification and

190

K. D. Cicerone

analysis of relevant attentional difficulties in their everyday functioning, and facilitating the application of within-session strategies to everyday functioning. While treatment was individually tailored to reflect participants’ clinical needs, there were several common themes. For example, initial treatment sessions were directed toward providing the participants with an interpretation of their attention deficits and a rationale for the intervention. The impact of slowed processing speed on performance was discussed, particularly the interrelationship between attention capacity and processing speed. The relationship between attention difficulties and subjective symptoms such as irritability and fatigue was explored. A rationale for the experience that `thinking was more effortful and less automatic’ was provided. Subsequent treatment sessions were directed at facilitating the participants’ conscious and deliberate use of strategies to effectively allocate attentional resources and manage the rate of information. Verbal mediation, rehearsal, anticipation of task demands, and self-pacing strategies were introduced and practiced by the treatment participants throughout the various training tasks. Self-paced trials were timed, and feedback was provided regarding the relationship between pace and errors during task performance, as well as the need to utilize different strategies during self-paced and therapist-paced trials. During self-paced, dual task trials the participants were occasionally able to pause one activity allowed to reinforce the self-management of time demands. Participants were often asked to self-monitor their èffort’ during performance, and this was related to task variations and objective indices of performance. Participants were encouraged and reinforced for maintaining task performance even when frustrated, and the use of positive self-statements for management of emotional reactions was practiced. Ìntrusive worrying’ by participants over their performance was interpreted as an additional cognitive demand which utilized limited resources, thereby contributing to attentional difficulties. During response generation and dual-task trials, participants were initially instructed to maintain their performance on the primary task while `sharing’ resources; on some occasions, however, participants were instructed to place priority on the `secondary’ task or to shift priorities between primary and secondary tasks within a trial. Participants frequently monitored the occurrence of àttention lapses’ during daily activities between treatment sessions, including the precipitating situation and use of strategies to regulate attentional resources and manage attentional demands. The later treatment sessions, especially during the dual-task trials, were increasingly devoted to fostering the participants’ self-appraisal and application of strategy use in the context of their everyday functioning. Overall, an integral aim of the treatment was to increase participants’ control over the allocation of attention resources, in order to improve their actual competence and perceived self-efficacy. None of the participants were receiving any other rehabilitation therapies during this period. Outcome measures Outcome measures were selected to reflect three levels of analysis. First, four common clinical tests of attention (yielding six measures) were administered to all participants before and after treatment. The Trailmaking Test Parts A and B were administered and scored according to the procedures and demographic norms provided by Heaton et al. [13]. The Paced Auditory Serial Addition Test (PASAT)


191

consisted of four sets of 50 single digits each, presented at interstimulus intervals of 2.4, 2.0, 1.6 and 1.2 seconds. The total number of correct responses on all four trials was selected for analysis. The total number of correct responses on the PASAT were transformed to age-related z-scores using the extended norms reported by Roman et al. [14] and then converted to t-scores. The Continuous Performance Test of Attention [15] total error raw scores were corrected for age and education and converted to t-scores. The 2 & 7 Test [16] was scored for Automatic Detection Speed and Controlled Processing Speed using age and education corrected t-scores. The effect of treatment on the clinical outcome measures was evaluated on the basis of clinically meaningful change following treatment. A pre- to post-treatment t-score difference greater than 1 SD (i.e. more than 10 points) was set, a priori, to represent `clinically significant change’, as this index has been used in previous studies of outcome from cognitive rehabilitation [17]. In addition, the reliability of clinically meaningful change was examined for individual participants, using the procedure described by Ingraham and Aiken [18] for determining overall abnormality on multiple measures. Based on this procedure, the criteria of clinically meaningful improvement on three or more of the six neuropsychological measures was required for each individual (p ˆ 0:05). In addition to the clinical measures, it was evaluated more specifically whether the intervention resulted in a change in specific components of attention, namely, processing speed and working memory efficiency. The longest string of consecutive correct responses on PASAT Trial 1 (2.4 seconds) and Trial 4 (1.2. seconds) were determined for each subject pre- and post-treatment. It was felt that this measure would reflect the participants’ ability to continuously update their mental representation of the previous digit in order to maintain efficient performance, thereby reflecting the control of auditory-verbal working memory [2]. The longest string on Trial 1 was considered to represent primarily this working memory component, while the longest string on Trial 4 would, in addition, reflect the influence of increased processing speed demands. A self-report measure was administered before and after treatment as a face valid measure of participants’ subjective experience of attentional difficulties in their everyday activities. Consideration was given to several existing rating scales, such as the Short Inventory of Minor Lapses (SIML) [19] and the Rating Scale of Attentional Behaviours [20]. However, the Rating Scale of Attentional Behaviours was developed for use with patients with severe brain injury during acute rehabilitation, and its psychometric properties have not been examined. Preliminary use of the SIML, which was developed for use with non-neurologic populations, suggested that it did not adequately capture the complaints of patients with MTBI. Therefore, a new self-report measure modelled after the SIML, the Attention Rating and Monitoring Scale (ARMS), was developed in an attempt to reflect the complaints of attention difficulty common to MTBI. The ARMS consists of 15 items relating to problems with concentration (e.g. `difficulty concentrating in noisy or busy situations’, `difficulty keeping your mind on an activity for more than a few minutes’), mental effort (e.g. `becoming very fatigued during activities where you have to pay attention’, `becoming easily overwhelmed due to difficulties paying attention’), and cognitive symptoms associated with attention difficulties (e.g. `difficulty returning to a task after being interrupted’, `forgetting things immediately after being told’). Participants rated how often they had experienced each difficulty in their day-to-day functioning over the past 2 weeks, using a five point scale

192

K. D. Cicerone

Table 1.

Neuropsychological measures of attention (t-scores) on initial evaluation (PRE) and re-evaluation (POST) for treatment and comparison participants 2&7 Speed

Trailmaking A

E1 E2 E3 E4 C1 C2 C3 C4

Trailmaking B

CPTA

PASAT

Automatic

Controlled

pre

post

pre

post

pre

post

pre

post

pre

post

pre

post

37 47 33 61 24 24 33 27

41 55 45* 57 24 41* 43 27

46 62 52 58 30 36 34 41

59* 66 62 65 34 49* 36 46

20 20 20 34 29 20 20 20

25 48* 52* 57* 20 20 25 20

42 37 43 56 38 37 44 24

57* 50* 57* 62 41 37 55* 29

35 41 42 49 42 49 28 34

41 52* 63* 66* 44 49 32 34

35 46 51 35 35 48 26 35

38 57* 69* 75* 43 48 30 34

E ˆ treatment participant, C ˆ comparison participant. * clinically meaningful change (pre± post t-score difference >10).

ranging from never (1) to always (5), and item scores were summed for a total possible score from 15± 75. Although the formal psychometric properties of the scale have not been examined, it has been shown to discriminate between a group of patients diagnosed with MTBI and a group of non-injured controls [6]. Results The treatment group exhibited significantly greater change on standard measures of attention, demonstrating clinically meaningful improvement on 58.3% of measures, while the comparison group demonstrated clinically meaningful change on 12.5% of measures (À2 ˆ 9:11, df ˆ 1, p ˆ 0:0025) (table 1). Examination of pre-treatment standard scores suggested that the treatment group was initially less impaired than the comparison group (À2 ˆ 5:58, df ˆ 1, p ˆ 0:018). In order to examine whether this difference in initial level of impairment contributed to the greater clinical change in the treatment group, the analysis was repeated using only initially impaired measures. The difference between groups remained significant (À2 ˆ 7:94, df ˆ 1, p ˆ 0:0048) with the treatment group demonstrating clinically meaningful improvement on 60.0% of initially impaired measures while the comparison group demonstrated improvement on 5.3% of initially impaired measures. In addition, there was no relationship between pre-treatment scores and change scores (Spearman rho ˆ 0.06), suggesting that this did not account for the greater change in the treatment condition. Analysis of clinically meaningful improvement for individual participants indicated that three of the four treatment participants demonstrated significant improvement, while none of the comparison participants did so. There were no differences between treatment and comparison groups, prior to treatment, on the PASAT Longest Strings correct for Trial 1 (16.0 vs. 11.3) or Trial 4 (3.5 vs. 3.0), indicating that the groups were initially equivalent on this index of working memory. The treatment group exhibited significantly greater improve-


193

ment following treatment on the Longest String correct on Trial 1 of the PASAT (30.0 vs. 9.75; Mann-Whitney U ˆ 16.0, p ˆ 0:021). On PASAT Trial 4, the degree of improvement between groups was not significant (12.0 vs. 5.3; MannWhitney U ˆ 3.0) suggesting a persistent impairment of processing speed. There was no difference between treatment and comparison groups in their initial self-report of attentional difficulties on the ARMS (59.3 vs. 61.6). Following treatment, the treatment group demonstrated significantly greater reduction in the experience of attentional difficulties (42.5 vs. 56.0; Mann-Whitney U ˆ 16, p ˆ 0:021) Discussion The results of this investigation support the effectiveness of a treatment for attention deficits after MTBI, based upon the conceptualization of deficits and interventions as a function of the central executive component of working memory, or `working attention’. Clinically meaningful improvement on attention measures was apparent as an effect of treatment, relative to an untreated comparison group. Further analysis of the PASAT suggested that the principal effect of the intervention was on the working memory component, that is the ability to temporarily maintain and manipulate information during task performance, with no direct effect on processing speed. Gray et al. [2] found a similar treatment effect on the PASAT when stimuli were presented at a relatively slow rate of one digit every 4 seconds. Despite this finding, improvement was noted on several clinical measures incorporating demands for speeded stimulus processing. This may be understand in relation to the participants’ increased awareness and management of temporal demands, rather than an improvement in discrete, second-to-second processing. Fasotti et al. [21] developed an intervention for `time pressure management’ to compensate for the consequences of slowed information after severe TBI. The integral aspect of this procedure was teaching participants management strategies which would give them enough time to manage tasks. These procedures included anticipating task demands, interrupting tasks, asking for clarification, and repeating information. The intervention was effective in improving participants performance, primarily through the increased use of managing steps to reduce time pressure (such as interrupting ongoing task demands). Of interest, the intervention also unexpectedly seemed to improve participants performance on the PASAT, despite the externally controlled, fixed rate of stimulus presentation on this task. The authors suggested that this improvement may have been due to the patients’ taking a `more assertive and less anxious attitude towards time pressure’ ([21], p. 63). In the present study, the intervention emphasized the participants’ conscious and deliberate use of strategies to effectively allocate attentional resources and manage the rate of information, which included verbal mediation, self-pacing strategies, sharing attentional resources during multiple tasks, self-monitoring of mental effort, and management of secondary emotional reactions during task performance. The results are, therefore, consistent with a strategy training model of remediation, in which the benefits of treatment are related to the participants improved ability to compensate for residual deficits and consciously adopt strategies for the more effective allocation of their remaining attentional resources [2, 21]. The present intervention also included an explicit attempt to foster the application of strategies outside of the immediate treatment situation. Following treatment,

194

K. D. Cicerone

the treatment participants showed a significant reduction in their complaints of attentional dysfunction in their daily lives. While not subject to formal analysis, it is worth noting that all four of the treatment participants resumed their vocational and social roles following treatment, while none of the comparison participants did so during this period. These findings suggest the potential for the generalization of treatment effects to participants’ everyday functioning. The interpretation of these results is limited by the small number of participants, and the lack of randomized assignment to the treatment condition. It should also be noted that the treatment participants were initially less impaired than the comparison group. The influence of initial severity of impairment on treatment effectiveness is not clear. Although more severely impaired patients may be more likely to demonstrate spontaneous recovery of attention and slowed processing [22], less severe initial impairment may be necessary in order for patients to acquire and apply compensatory strategies. In the present study, it did not appear that the less severe initial impairment in the treatment group accounted for their greater improvement, and Gray et al. [2] did not detect a differential treatment effect between participants with mild± moderate versus severe initial impairments. In sum, intervention based on a cognitive neuropsychological model of working memory may be effective in the remediation of attention deficits after TBI. While this approach appears to hold promise, it is worth reinforcing the view that `the therapy is not synonymous with the task’ [23]. For example, in the current intervention the n-back tasks were chosen and designed, on clinical and theoretical grounds, primarily as a means of eliciting the patient’s deficits within the treatment situation. The tasks also provided a cogent, working model for interpreting the patient’s performance and subjective experience, both within treatment and in their daily functioning. There was no expectation that the patient would derive any therapeutic benefit simply from being exposed to and practicing these tasks. In fact, the therapeutic intervention required intensive patient-therapist interaction and participation in activities beyond the specific tasks, including cognitive selfmonitoring, emotional self-appraisal, and the imagined use of strategies in real situations. These procedures, within the context of a strategy training model of neuropsychological rehabilitation, merit further, controlled investigation. The contribution of patient-based moderating variables, such as initial level of impairment, also requires additional consideration. References 1. Mateer, C., Sohlberg, M. M. and Youngman, P. K.: The management of acquired attention and memory deficits. In: R. Ll. Wood and I. Fussey (editors) Cognitive Rehabilitation in Perspective (New York: Taylor & Francis), 68± 95, 1990. 2. Gray, J. M., Robertson, I., Pentland, B. et al.: Microcomputer-based attentional retraining after brain damage: A randomized group controlled trial. Neuropsychological Rehabilitation, 2: 97± 115, 1992. 3. Niemann, H., Ruff, R. M. and Baser, C. A .: Computer assisted attention retraining in head injured individuals: a controlled efficacy study of an out-patient program. Journal of Consulting and Clinical Psychology, 58: 811± 817, 1990. 4. Ponsford, J. L. and Kinsella, G.: Evaluation of a remedial programme for attentional deficits following closed head injury. Journal of Clinical and Experimental Neuropsychology, 10: 693± 708, 1988. 5. Sturm, W., Wilmes, K. and Orgass, B.: Do specific attention deficits need specific training? Neuropsychological Rehabilitation, 7: 81± 103, 1997.


195

6. Cicerone, K. D., Azulay, J. and Kalmar, K.: The nature of attention impairment after mild traumatic brain injury. Submitted for publication, 2000. 7. Cicerone, K. D.: Attention deficits and dual task demands after mild traumatic brain injury. Brain Injury, 10: 79± 89, 1996. 8. D’ Esposito, M., Detre, J. A ., Alsop, D. C. et al.: The neural basis of the central executive system of working memory. Nature, 378: 279± 281, 1995. 9. McAllister, T. W., Saykin, A. J., Flashman, L. A. et al.: Brain activation during working memory 1 month after mild traumatic brain injury: A functional MRI study. Neurology, 53: 1300± 1308, 2000. 10. Baddeley, A. and Hitch, G. J.: Developments in the concept of working memory. Neuropsychology, 8: 485± 493, 1994. 11. Baddeley, A.: Working memory or working attention? In: A. Baddeley and L. Weiskrantz (editors) Attention: Selection, awareness & control (Oxford: Clarendon Press), pp. 152± 170, 1993. 12. Azouvi, P., Jokic, C., Van Der Linden, M. et al.: Working memory and supervisory control after severe closed-head injury. A study of dual task performance and random generation. Journal of Clinical and Experimental Neuropsychology, 18: 317± 337, 1996. 13. Heaton, R. K., Grant, I. and Matthews, C. G.: Comprehensive norms for an expanded HalsteadReitan battery (Odessa, FL: Psychological Assessment Resources, Inc), 1991. 14. Roman, D. R., Edwall, G. E., Buchanan, R. J. et al._: Extended norms for the Paced Auditory Serial Addition Test. The Clinical Neuropsychologist, 5: 33± 40, 1991. 15. Cicerone, K. D.: Clinical sensitivity of four measures of attention to mild traumatic brain injury. The Clinical Neuropsychologist, 11: 266± 272, 1997. 16. Ruff, R. M. and Allen, C. C.: Ruff 2 & 7 Selective Attention Test Professional Manual (Odessa, FL: PAR), 1996. 17. Rattok, J., Ben-Yishay, Y., Ezrachi, O. et al.: Outcome of different treatment mixes in a multidimensional neuropsychological rehabilitation program. Neuropsychology, 6: 395± 415, 1992. 18. Ingraham, L. J. and Aiken, C. B.: An empirical approach to determining criteria for abnormality in test batteries with multiple measures. Neuropsychology, 10: 120± 124, 1996. 19. Reason, J.: Self-report questionnaires in cognitive psychology: have they delivered the goods? In: A. Baddeley and L. Weiskrantz (editors) Attention: Selection, Awareness and Control (Oxford: Clarendon Press), 406± 423, 1993. 20. Ponsford, J. L. and Kinsella, G.: Evaluation of a remedial programme for attentional deficits following closed head injury. Journal of Clinical and Experimental Neuropsychology, 10: 693± 708, 1988. 21. Fasotti, L., Kovacs, F., Eling, P. A. T. M. et al.: Time pressure management as a compensatory training strategy after closed head injury. Neuropsychological Rehabilitation, 10: 47± 65, 2000. 22. Spikman, J. M, Timmerman, M. E., van Zomeran, A. H. et al.: Recovery versus retest effects in attention after closed head injury. Journal of Clinical and Experimental Neuropsychology, 21: 585± 605, 1999. 23. Basso, A. and Marangolo, P.: Cognitive neuropsychological rehabilitation: the emperor’s new clothes? Neuropsychological Rehabilitation, 10: 219± 229, 2000.