Deficits in attention are more commonly found among individuals with more severe traumatic brain injuries (TBI), and may encompass delayed reaction time, reduced speed of information processing, or challenges with concentration, forgetfulness, or doing more than one thing at a time (e.g., walking and talking). This chapter presents cognitive rehabilitation therapy (CRT) interventions aimed to restore attentional capacity, divided by phase of recovery following moderate-severe TBI (i.e., subacute and chronic). Controlled studies are described in detail within these sections, divided by treatment comparator arm, followed by descriptions of the noncontrolled studies. The committee’s conclusions are presented at the end of the chapter.
The committee reviewed six randomized controlled trials (RCTs), including two crossover studies, of treatments intended to improve attention. All six involved modular treatment directed at one or more attentional processes. All used decontextualized treatment materials, and all were categorized as restorative. The trials involved a total of 264 study participants; treatment group sizes in individual trials ranged from 7 to 43 patients. Nearly all of the patients suffered moderate-severe injuries 6 weeks to many months prior to study enrollment. Study participants were generally in their late 20s to early 30s.
The committee did not identify any nonrandomized, controlled parallel group designs of treatments for attention deficits, however it did review two pre-post single group studies and one single-subject, multiple baseline experiment. These studies also employed primarily modular restorative
treatments, and all were delivered to patients in the chronic phase with moderate-severe injuries. The committee did not identify any studies assessing CRT interventions for attention in patients with mild TBI. Table 7-1 presents a summary of all included studies in this review.
Subacute Phase of Recovery
Comparator Group: Non-CRT Content
Gray et al. (1992) compared approximately 17 hours of computer administered modules stressing various dimensions of attention to about 12 hours of recreational computing that excluded externally paced tasks or tasks that required rapid processing and responding. This study found a positive effect of training on psychometric measures of attention, particularly the type that require numerical manipulation in working memory. These effects grew in significance in follow-up compared to the immediate posttreatment measures. This pattern is of some concern, since the median time postinjury was 20 weeks, a point at which natural recovery may be ongoing; therefore, imbalance in the acuity of injury between groups might produce such a result. However, time postinjury was statistically controlled for, and measures of functions unrelated to attention did not show greater improvement in the treatment group, lending some specificity to the findings. In this study nearly half of the subjects had nontraumatic injuries, but the authors report no interaction between diagnosis and treatment benefit. The credibility of this study is compromised due to its nonreporting of sample sizes for analysis posttreatment, especially at the 6-month follow-up. Furthermore, standard deviations of the outcomes were not provided.
Comparator Group: Other CRT Content
Novack et al. (1996) studied participants who were 3 to 6 months postinjury. This study was conducted in an acute inpatient rehabilitation population approximately 3 to 6 weeks postinjury, a time when many of the patients were confused and highly impaired. One group received a structured program of attention training. The other group received a variety of other rehabilitation interventions that involved cognitive rehabilitation components that did not specifically focus on attention. Outcomes were assessed with respect to several psychometric measures of attention as well as the Functional Independence Measure (FIM). Both groups improved significantly from pre- to posttreatment, but to a comparable degree.
TABLE 7-1 Evidence Table: Attention
Study | N | TBI Severity Level | Brief Narrative | Comparator | Outcome Measures | Findings |
RCT Gray et al. 1992 |
31 | Mild, moderate, severe |
Patients with attentional deficits randomized to two groups. The treatment group received computerized attention training including: • Reaction time training • Rapid number comparison • Digit symbol transfer • Alternating Stroop program • Divided attention tasks |
Y Non-CRT Content: Recreational computing |
• Psychometric measures: ■ Paced Auditory Serial Addition Test (PASAT) ■ Wechsler Adult Intelligence Scale-Revised (WAIS-R) • Frontal functions: ■ Wisconsin Card Sorting Test (WCST) ■ Finger tapping ■ Word fluency • Other attentional functions: ■ Letter cancellation ■ Picture completion ■ Time estimation |
Significant differences between groups on psychometric measures of attention at immediate and six-month follow-up. |
McMillan et al. 2002 | 145 | Moderate | Sessions of supervised mindfulness practice using an audiotape obtained from Jon Kabat-Zinn. Participants were asked to practice daily with this tape in the intervening periods. |
Y No or Non-CRT Content: • Group with therapist contact and physical exercise (PE) • Group with no treatment |
• Test of Everyday Attention • Adult Memory and Information Processing Battery • PASAT • Trail Making Test • Self-reported: ■ Sunderland Memory Questionnaire ■ Cognitive Failures Questionnaire ■ Hospital Anxiety and Depression Questionnaire |
Between-group comparison not significant for any measure. |
Niemann et al. 1990 | 26 | Moderate-Severe | Computerized attention training with visual, auditory, and divided (i.e., both) training components, further subdivided into focused and alternating attention tasks Training accompanied by feedback and strategy teaching | Y Other CRT Content: Memory training of equivalent duration and intensity |
■ General Health ■ Rivermead Post-Concussional Symptoms Questionnaire • Attention ■ Test d2 ■ PASAT-R ■ Divided Attention Test ■ Trail Making Test, Part B • Memory ■ Rey Auditory Verbal Learning Test-Modified (RAVLT-M) ■ Block Span Learning Test |
Authors report attention group improved significantly more than control on the three measures of attention; memory group improved more than attention group on one measure. |
Novack et al. 1996 | 44 | Severe | Hierarchical training of attention skills for lowest level of functioning involved focused and sustained attention, more challenging tasks requiring selective attention, alternating attention, and divided attention. Each level of the hierarchy included multiple tasks. | Y Other CRT Content: Unstructured stimulation program |
• Wechsler Memory Scale-Revised (WMS-R) • Single reaction time • Choice reaction time • Functional Independence Measure (FIM).Activities of Daily Living • FIM.Cognition • Neuropsychological tests: ■ Logical Memory I and II ■ Sentence Repetition Test |
Between-group comparison not significant for any measure. |
■ Judgment of Line Orientation ■ Trail Making Test ■ Wide-Range Achievement Test-Revised (WRAT-R), arithmetic subtest ■ Visual imperception |
||||||
Ruff et al. 1994 | 15 | Severe | Two groups received two treatments (attention and memory training) delivered via THINKable, a computer-based, multimedia program. | Y Other CRT Content: Group A received attention training, then memory training. Group B received memory training, then attention training. |
• Attention ■ Ruff 2 & 7 Selective Attention test ■ WAIS-R Digit Symbol ■ Continuous Performance Test • Memory ■ Rey Auditory Verbal Learning Test ■ Corsi Block Learning Test • Behavioral assessments |
Attention measures at posttreatment did not show significant between-group comparison. |
Sohlberg et al. 2000 | 14 | Moderate-Severe | Compares Attention Process Training (APT) to education, delivered to two groups in a crossover design. APT aims to improve memory, learning, and some aspects of executive control. | Y No or Non-CRT Content: Placebo—Brain Injury Education and Supportive Listening |
• Neuropsychological attention tests: ■ Trail Making Test ■ PASAT ■ Gordon Diagnostic ■ Controlled Oral Word Association Task (COWAT) ■ Covert Orienting ■ Continuous Performance Task ■ Stroop task ■ Sternberg tasks • Self-report: ■ Attention Questionnaire ■ Brock Adaptive Functioning Questionnaire (BAFQ) ■ Dysexecutive Questionnaire (DEX) |
Training showed improvement on PASAT, Stroop, and Trail Making Test. Patients reported a significant difference in attention and memory in the treatment group versus education. Subjects with greater changes had improved score on PASAT. Significant more reports of improvement to daily life (via questionnaires) were given by participants after APT. |
Pre-Post Single Group Park et al. 1999 |
23 | Moderate-Severe | Examined the use of the Attention Process Training (APT) to determine if APT improves cognitive function versus learning specific skills. | N |
• PASAT • Consonant trigrams (Brown-Peterson task) • Beck Depression Inventory (BDI) |
Post-testing showed significant improvement on PASAT and Consonant trigrams. No significant improvement on BDI. Results are mediated by lack of control group. |
Stathopoulou and Lubar 2004 | 5 | Severe | Included a “Captain’s Log” computer training program for attention skills with tasks for vigilance, inattention, prudence, impulsivity, focus, variability, and speed. Participants selected, discriminated, or matched visual pictures or sounds. | N |
• Electroencephalogram (EEG) variables • WAIS-R • PASAT • Self-report measure of severity (of deficit) for memory and attention symptoms |
Sustained attention, alternating attention, and divided attention improved in three of five subjects. Selective attention improved in all subjects. Focused attention improved in two of five subjects. Results are mediated by lack of control group. |
Single Subject, Multiple Baseline Gansler and McCaffrey 1991 |
4 | Severe | Hierarchically organized attention training program based on Posner’s four component model of attention. | N |
• Neuropsychological tests: • WMS-R • Trail Making Test • Minute Estimation Test • Thurstone Word Fluency Test • Grooved Pegboard Test • WAIS-R • Psychological variables for depression, anxiety, and anger • Self-report of ADLs |
Posttreatment test results proved not significant for all participants. Participants reported increase in ADLs. |
Chronic Phase of Recovery
Studies of chronic, moderate-severe TBI included four RCTs (McMillan et al. 2002; Niemann et al. 1990; Ruff et al. 1994; Sohlberg et al. 2000) comparing five treatment arms with patients in the chronic phase. Interventions in three (Niemann et al. 1990; Ruff et al. 1994; Sohlberg et al. 2000) of these RCTs consisted of some form of attention training exercises, similar to those employed by Gray et al. (1992) (see above), and most were delivered via computer. Training ranged from 10 to 24 hours and typically involved several different attention-demanding tasks that progressed in difficulty with patient improvement. Some treatments included therapist-delivered goal setting, feedback, and review of performance, including one study of Attention Process Training (APT), a manualized treatment approach that specifies therapist feedback more systematically. The fourth RCT (McMillan et al. 2002), also the largest trial, used mindfulness training. Unlike the other attention treatments, mindfulness training did not involve practice with attention-demanding tasks but rather separate sessions focused on breathing. Therapist-led training in this study was fewer than 4 hours for both mindfulness training and the active comparison condition, but with home practice assigned.
Comparator Group: No or Non-CRT Content
McMillan et al. (2002) compared the effects of instruction in mindfulness training to comparable instruction in physical exercise (non-CRT content) and a no-treatment control where participants received no therapist contact but were assessed at the same intervals. Thus, this was the only study that had a comparator arm of no treatment. Outcomes were assessed in terms of neuropsychological measures of attention as well as several self-report measures of mental health status and lapses of attention in everyday life. The mindfulness intervention outcomes on attention were no different than those of physical exercise or no intervention.
Sohlberg et al. (2000) compared 24 hours of manualized APT delivered over 10 weeks to 10 hours of brain injury education—a non-CRT intervention—delivered over the same time period, in an RCT with outcomes assessed at the point of crossover and again at trial completion. Outcome measures included standardized neuropsychological measures of attention, laboratory measures of information processing intended to assess the functioning of specific neural networks subserving separable attentional domains, and coded qualitative interviews regarding real-world changes resulting from treatment. This trial found positive effects of attention training on the Paced Auditory Serial Addition Test (PASAT), a measure of working
memory and speeded mental addition, and on the Memory for Location task, a measure of location working memory. On the Stroop task and the Trail Making Test, members of the APT group were characterized by “low vigilance” at baseline. The trial did not find such effects on verbal working memory, verbal fluency, or on the laboratory tasks designed to isolate the functions of specific neural networks. Although the patients were not blinded to the content of their treatment, there were significantly more reports of attention improvements in daily life after the APT treatment than after brain injury education. Lending some support to the validity of these reports, reports of everyday attention benefits correlated with improvement in PASAT scores. This was a small study, with 14 participants, all with moderate-severe injuries. Two subjects were not included in the structured interview to assess improvement because they did not recall their participation in the treatment. This situation is problematic, as it reduces the sample size to 12 and raises concerns about generalization to patients with substantial memory impairment. In addition, there were several statistical tests, with no adjustment for multiple testing.
Comparator Group: Other CRT Content
Two trials (Neimann et al. 1990; Ruff et al. 1994) studied the impact of an attention training program, compared to a memory training program, on measures of attention; thus memory training served as the control treatment.
Neimann et al. (1990) provided approximately 36 hours of training on three different aspects of attention, or a comparable amount of training on internal and external memory strategies. Neuropsychological measures of attention and memory were assessed. Based on a significant result from a MANOVA test for the four attention measures, the authors reported “partial support” for the treatment prediction that attention training would provide more robust impact on attention measures than the comparison memory training. However, in post hoc testing, only one of the attention measures differed significantly between groups. Inspection of the pattern of improvement suggests that three attention measures improved more in the group that received attention training, and one improved more in the group that received memory training.
Ruff et al. (1994) conducted a similar study in which the two treatment groups received both attention training and memory training, but in counterbalanced order. However, the authors did not conduct statistical testing at the midpoint of treatment (when a parallel group comparison would have been possible) because of the small sample size. They report benefit in both domains at the end of combined treatment, but inspection of
the pattern of scores at the midpoint suggests that some attention measures improved more in one group and some in the other.
Pre-Post Designs
Park et al. (1999) studied the effects of 40 hours of APT training in 23 individuals with chronic, moderate-severe TBI using the PASAT and Consonant Trigrams tests as outcome measures, along with the Beck Depression inventory. Stathopoulou and Lubar (2004) studied five people with severe brain injury between 1.5 and 23 years postinjury. The patients received 18 hours of attention training using “Captain’s Log,” a commercial computerized product that administers tasks involving various challenges to verbal and visual attention and memory. Participants were tested only once pre and once post, using digit span and digit symbol subtests of the Wechsler Adult Intelligence Scale (WAIS), the PASAT, a continuous performance test, a self-report measure of severity of a number of attention and memory symptoms rated on a five-point scale from “no problem” to “severe problem,” and electroencephalogram (EEG) spectral measures. These studies—all of which were conducted at a time when rapid natural recovery would be unexpected—showed improvement in some of the outcome measures relevant to treatment. However, none of these studies had an adequate control for practice on the outcome assessments themselves, which were assessed twice, so none provides strong support for a treatment effect.
Single-Subject, Multiple Baseline Experiment
Gansler and McCaffrey (1991) conducted four single-subject experiments in which individuals with severe TBI—4 to 27 years postinjury—received repeated testing on a set of information processing measures modeled on Posner’s attention components. The measures were administered weekly, beginning 4 weeks prior to training, during the 8 weeks of training, and at 1 month after training. Training consisted of 8 weeks of hierarchically organized modules of attention totaling about 64 hours. Other psychological measures were also administered weekly and neuropsychological measures at baseline, after training, and at follow-up; participants also completed a self-assessment of ADL performance and their satisfaction with it. Improvement on attention measures and psychological measures was negligible for all participants, though there were larger effects on self-appraisal of ADL performance. This result could suggest that the treatment imparted compensatory skills for managing attention deficits that were evident in real-world ADL tasks but not on controlled attention processing tasks.
However, the result is also consistent with biasing of self-reported benefit because of expectation.
The committee found limited evidence from one RCT (Sohlberg et al. 2000) to support conclusions about the impact on patient-centered outcomes (quality of life, functional status) in moderate-severe TBI.
The committee found limited evidence from one RCT (Gray et al. 1992) on long-term impact of treatment (6 months) in the subacute phase as assessed with psychometric measures, particularly the type requiring numerical manipulation in working memory.
Considering subacute and chronic studies together, the committee found limited evidence from two studies (Grey et al. 1992; Sohlberg et al. 2000), that intensive practice of hierarchical attention-demanding tasks had a positive impact on psycho-metric measures of attention in the immediate posttreatment period and/or at follow-up.
The review did not include any RCTs or other study designs on CRT for attention in mild TBI. Two studies (Gray et al. 1992; Novack et al. 1996) provided limited evidence to conclude that CRT improves attention in subacute, moderate-severe TBI patients. In studies of moderate-severe TBI patients in the chronic phase of recovery, a few, relatively small RCTs with several methodologic limitations provided mixed support for treatment benefit. These trials tested intensive practice of hierarchical attention-demanding tasks on some psychometric measures of attention, with positive immediate outcomes. However, none studied the durability of benefits, and only one study assessed treatment impact with respect to patient-centered outcomes (i.e., Sohlberg et al. [2000] found a preliminary association of improved psychometric measures of attention with real-world benefits). Data from pre-post designs, although consistent with some treatment benefit, provide weak support because of the possible confounding effect of practice on the outcome measures.
Several of the RCTs with equivocal results (Niemann et al. 1990; Ruff et al. 1994) used intensive memory training as a control condition. Since all tasks requiring effort place demands on attention, it is possible that the overlap in treatment outcomes between treatment groups in such studies reflects the overlap in mental demands of treatment content, potentially attenuating or accounting for the lack of finding of differences in attention outcomes. Of note, the two studies that provided the strongest support for
the efficacy of hierarchical attention training employed non-CRT comparator conditions.
Gansler, D. A., and R. J. McCaffrey. 1991. Remediation of chronic attention deficits in traumatic brain-injured patients. Archives of Clinical Neuropsychology 6(4):335–353.
Gray, J. M., I. Robertson, B. Pentland, and S. Anderson. 1992. Microcomputer-based attentional retraining after brain damage: A randomised group controlled trial. Neuropsychological Rehabilitation 2(2):97–115.
McMillan, T., I. H. Robertson, D. Brock, and L. Chorlton. 2002. Brief mindfulness training for attentional problems after traumatic brain injury: A randomised control treatment trial. Neuropsychological Rehabilitation 12(2):117–125.
Niemann, H., R. M. Ruff, and C. A. Baser. 1990. Computer-assisted attention retraining in head-injured individuals: A controlled efficacy study of an outpatient program. Journal of Consulting and Clinical Psychology 58(6):811–817.
Novack, T. A., S. G. Caldwell, L. W. Duke, T. F. Bergquist, and R. J. Gage. 1996. Focused versus unstructured intervention for attention deficits after traumatic brain injury. Journal of Head Trauma Rehabilitation 11(3):52–60.
Park, N. W., G. B. Proulx, and W. M. Towers. 1999. Evaluation of the attention process training programme. Neuropsychological Rehabilitation 9(2):135–154.
Ruff, R. M., R. Mahaffey, J. Engel, C. Farrow, D. Cox, and P. Karzmark. 1994. Efficacy study of THINKable in the attention and memory retraining of traumatically head-injured patients. Brain Injury 8(1):3–14.
Sohlberg, M. M., K. A. McLaughlin, A. Pavese, A. Heidrich, and M. I. Posner. 2000. Evaluation of attention process training and brain injury education in persons with acquired brain injury. Journal of Clinical and Experimental Neuropsychology 22(5):656–676.
Stathopoulou, S., and J. F. Lubar. 2004. EEG changes in traumatic brain injured patients after cognitive rehabilitation. Journal of Neurotherapy 8(2):21–51.