Retraining Neural Pathways Improves Cognitive Skills After A Mild Traumatic Brain Injury

University of California, San Diego, University of South Alabama, University of California, Riverside

The proposed study tests the feasibility (Phase I) of PATH neurotraining to improve working memory and attention in mTBI patients rapidly and effectively to provide clinical testing of a therapeutic training for the remediation of cognitive disorders caused by a concussion. This study will contribute to the fundamental knowledge of how to remediate concussions from a mTBI to enhance the health, lengthen the life and reduce the disabilities that result from a mTBI.

This study will provide clinical testing of therapeutic training for the remediation of cognitive disorders caused by a concussion. A Phase I Clinical Trial will be conducted to determine the feasibility of targeting an etiology of visual timing deficits in mTBI subjects to be successful in treating cognitive deficits caused by the mTBI. This clinical trial will increase not only our understanding of the range of cognitive deficits incurred after a mTBI, but also the best intervention to treat cognitive deficits after a mTBI. As soon as study protocol is approved by an IRB, brochures will be posted at both Kaizen Brain Center by Dr. Ahmed and the VA hospital by Dr. Huang to recruit subjects. Interested subjects will be screened by phone to ascertain their eligibility for MEG/MRI study. This study is designed to provide a clearer understanding of, and training targeted toward cortical timing processes (PATH training). The investigators will extend previous results from a pilot study of 4 mTBI subjects to a much larger sample of mTBI subjects. The proposed study tests the feasibility (Phase I) of PATH neurotraining to improve processing speed and visual working memory (primary outcomes) and reading speed, auditory WM, selective attention, cognitive flexibility, sustained attention, reading proficiency, post-concussive symptoms, quality of life and MoCA (secondary outcomes) in mTBI subjects rapidly and effectively, by analyzing the difference in scores on standardized tests before and after intervention training. This study will compare PATH training, presenting dim gray patterns moving left or right to activate the dorsal stream, with Sham training, presenting high contrast colored stationary patterns tilted left or right to activate the parvocells in the ventral stream, with an N-Back Working Memory (WM) task to primarily activate dorsal lateral PreFrontal Cortex (dlPFC), as done by competition. The investigators predict that PATH training improves working memory, processing speed and attention more than Sham or N-Back WM training after a mTBI. Since PATH training must be followed by cognitive exercises to improve cognitive function, a complementary strategy consisting of 15 minutes of working memory practice, recalling the correct sequence of digits, each presented for 500 msec, from 5 digits up to 10 digits will be completed for 15 minutes following both PATH and Sham training. Three interventions (PATH, N-back WM, and Sham neurotraining) each 30 minutes, will be administered three times/week for 12 weeks. The individuals entering into this study will come from general practice in Neurology, and may therefore include individuals involved in athletic injuries, motor vehicle accidents, and work-related injuries, as well as from the VA Hospital having a combat-related mTBI. As such the location and extent of brain trauma can be expected to be relatively heterogeneous. Inclusion/Exclusion criteria are described elsewhere. This study targets 80 mTBI adult subjects between the ages of 18-55 to conduct all three Specific Aims, using a single coordinated test site (UCSD-PDI). The treatment and control groups will be balanced in terms of age, sex, visual working memory, Montreal Cognitive Assessment (MoCA) scores, number of TBI incidents, duration of loss of consciousness, and locus and extent of the mTBI. Approximately 40 subjects, 13 subjects in each of three training groups will be studied each year for a total of 2 years. Half of the subjects in the treatment groups will have pre-post MEG exams to provide biomarkers of improvements in cognitive skills. There will be a rolling enrollment of subjects with at most 15 subjects at one time, enabling subjects to be monitored easily by UCSD staff. A key connecting actual subject name and ID number in stored data will be kept in a locked file cabinet and stored on a computer with password protection so subject confidentiality is preserved. Since patient burden is high by being trained 3 times a week for 12 weeks, subject compensation and desire for improved cognition should encourage mTBI subjects to remain in study. Less than a 5% drop-out rate is expected, since subjects will not be compensated for participating until they have completed this study. The statistician will randomly assign each subject to one of the three intervention training groups, controlling for all factors listed above. This study incorporates a parallel design to investigate which training group improves the most. Subjects will be studied over two time periods, 12 weeks apart (pre-post testing). Intervention training will be implemented in a high-fidelity manner using: 1) a detailed written protocol that all Research Assistants (RAs) are trained to follow meticulously, 2) training videos, and 3) having all RAs complete PATH, N-Back WM, and Sham interventions before administering this training to subjects. To ensure all tests are administered the same way, written instructions are said verbatim, and all training is supervised by the PI and/or SRA. UCSD students will be recruited at UCSD under the co-sponsorship of Professor Rossano. Intervention training will be conducted at either PDI's office in Solana Beach or at the Radiology Imaging Lab at UCSD, choosing the location that is more convenient for the subject. A second SRA will oversee all the MEG data collection at the Radiology Imaging Lab at UCSD, coordinating activities with the behavioral SRA at PDI. Training will be done during the day at a convenient time for the subject to do for 30 minutes 3 days/week for 12 weeks. New PATH training videos aimed at mTBI subjects will be developed, so are maximally relevant to adults with mTBI. The RAs will: 1) follow the study protocol by making sure the subject is focused on task, 2) be trained extensively to administer the standardized tests, and 3) being blind to which mTBI subjects are in which group when administering standardized tests. Learning to administer the standardized tests and treatment and control interventions will be done with the PI and SRA supervising each team of new study RAs. Each subject will be trained one-on-one to ensure task is done correctly. RAs ensure the subject is on task by examining computer data, and provide more training when needed. All subjects must do supervised training to ensure training is done consistently, at same time of day, for 30 min 3 times/week for 12 weeks. Otherwise, completing this task in a timely manner is not likely to occur, especially since mTBI subjects are learning to pay attention and remember more easily. Text reminders will be sent to each subject, as well, to remind them to come in for the brain training. If a subject is experiencing some discomfort, they will be told to look away from screen, and take a short 1-2 minute break. These side effects are rarely experienced. The PI, Dr. Teri Lawton, and the Co-I, Prof. Aaron Seitz, both having extensive experience conducting controlled validation studies, and the 2 SRAs (one overseeing behavioral data and one MEG-based data) will be in charge of training all staff, running daily operations which requires scheduling and supervising: 1) pre-post neuropsych testing, 2) pre-post MEG exams at UCSD, and 3) administering the treatment and control interventions. The PI cannot influence the results, since data is collected by the SRA or RA when neuropsychological tests are administered, or automatically by all intervention testing programs, and behavioral data will only be analyzed by statistician. PDI's SRA will enter all data into the REDCap database to be analyzed by statistician who will enter data into FITBIR database. Studies with a biomarker component will be used for half subjects in each treatment group to determine the pre-post timing and functional capabilities of different cortical areas in the attention and executive control pathways, in conjunction with behavioral pre-post standardized tests of cognitive abilities. The treatments being examined promote sustained functional recovery from a mTBI, including interventions during the chronic phase of injury, when currently there are no proven solutions for mTBI (80% of TBIs). MEG recordings before and after training will provide a neural correlate biomarker to determine whether PATH training improves the function of the dorsal, attention, and working memory networks more than found after a sham treatment and after a N-Back WM treatment. MEG recordings will also be used to determine whether PATH training strengthens coupled theta/ gamma activity, and/or alpha/ gamma activity. To increase its commercialization ability, PATH neurotraining must be shown to improve brain function using a biomarker, as stated by neurologists and therapists in letters of support. A structural MRI used for superimposing the functional activity on top of the brain anatomy will be collected before initial MEG recording. To evaluate effectiveness of PATH, and WM neurotraining, the magnitude of improvements in MEG physiological recordings in the first 100-200 msec interval will be used to examine the functional changes in the visual system, and the 200-1000 msec interval will be used to study the changes in the later responses of the WM network. Professor Huang and his staff will record voxel-wise MEG source magnitude images, physiological recordings from half of the mTBI subjects in the treatment groups (using a neural correlate biomarker to show feasibility of treatments to improve cognitive ability). MEG images covering the whole brain, and each frequency band, following the Fast-VESTAL procedure to measure time-locked signals during an N-back WM task, will be used to evaluate improvements in brain function, as done in a pilot study. Three interventions (PATH, N-back WM, and Sham neurotraining) each 30 minutes, will be administered three times/week for 12 weeks. The investigators will examine moderators that may determine training outcome. Individual differences at initial assessment will be evaluated to determine whether they predict improvements following training for different subpopulations: 1) Veteran mTBI subjects vs. civilian mTBI subjects, 2) different age groups (18-28, 29-41, 42-55), and 3) different loci and extent of the mTBI deficit. Whether these timing-based deficits predict cognitive skill deficits that are moderated by individual factors such as number of concussions, duration of loss of consciousness (LOC), nature and extent of cognitive deficits, cause of mTBI (blast or car accident or fall, for example), and age will be examined. These assessments will help to better understand the effects of a mTBI on different loci of mTBI deficit. Studies with a biomarker component will be used to determine the pre-post timing and functional capabilities of different cortical areas in the attention and executive control pathways, in conjunction with behavioral pre-post standardized tests of cognitive abilities. This exploratory aim will be used to determine whether different treatments should be explored during Phase II for different types of mTBI and different ages. The purpose of this study is to test the feasibility of PATH training to improve neural timing and cognitive skills after a mTBI. We use the most precise methods to measure improvements in cognitive skills for mTBI subjects pre- and post- training by using established standardized tests of cognitive skills that will be analyzed by the statistician using mixed-factors Multiple Analysis of CoVariance (MANCOVA) and Multi Level Modeling (MLM) statistics. The proposed statistical analyses are more than sufficient to answer the questions being examined using a well-planned statistical design. Bi-weekly meetings of our team will be conducted on zoom to ensure this rigorous study is on track, and to prepare results for scientific meetings and publishing in high-impact scientific journals. Whether cognitive improvements are sustained over time will be evaluated during Phase II.

The investigators will show that working memory, processing speed, reading speed, and attention improve more in mTBI patients after neurotraining that improves sensory and processing speed functions, improving the motion working range (PATH), more than after Sham training that activates a different visual pathway, or N-Back WM training that activates dlPFC to show the feasibility of PATH training to improve cognitive skills the most. MEG brain imaging will be used to show that PATH neurotraining improves the function of the dorsal, attention, and working memory networks (e.g., PPC, DLPFC and ACC/PCC areas), reducing over recruitment, more than found after N-Back WM training.

Staff will not be told whether the subject was in the treatment or control group. The statistician will not know which group was the treatment group and which was the control group.

Subject looks at computer screen to determine whether dim gray stripes in fish-shaped window move left or right relative to stationary background stripes. The subject reports which way center stripes move by pushing left or right arrow key, receiving brief tone if incorrect. Program adaptively changes contrast of test pattern in order to keep subject at 79% correct. There are levels of difficulty introduced by making the background pattern more similar to that in fish, by increasing pattern's complexity level, and by increasing number of directions of movement from one to two directions of motion. Intervention will be trained for one training cycle, 15 minutes, 3 times each week for 16 weeks. Fifteen minutes of working memory practice, recalling the correct sequence of digits, each presented for 500 msec, from 5 digits up to 10 digits will be completed for 15 minutes following PATH training.

The sham treatment will be Orientation Discrimination training that is identical to PATH training except instead of low contrast sinewave gratings moving left or right, 100% contrast stationary test and background sinewave gratings are used, both red, green, and black and white gratings, see patterns in Fig. 4 below. These patterns are randomly oriented left or right, at decreasing tilt angles as the test grating's orientation is identified correctly. These patterns only activate parvocells in ventral pathways (Ungerleider & Mishkin, 1982; Kaplan & Shapley, 1986) instead of activating dorsal pathways, the key component of PATH neurotraining. Therefore, this task does not speed up the brain's visual timing, which is a function of the dorsal stream. For the Orientation Discrimination task, the subject pushes the left arrow key when the test pattern is tilted left and the right arrow key when pattern is tilted right. Otherwise the two training tasks use the same paradigm.

Participants are required to compare each item on a computer screen to the item that they saw n-items back in the sequence. The participant plays a simple game where they control the movement of an astronaut that needs to collect correct gems, avoid incorrect gems, and also obstacles, to succeed at the game. If the participants performs well then they can be advanced to the 2-back, 3-back, 4-back, etc where they make similar matches but to earlier items in the sequence. This gamified task consists of a color n-back with 6 colors (a new color every 3 seconds, requiring subjects respond to targets by tapping the screen and navigating the astronaut to the targets, and avoiding the distractors) with 30% targets, and where the n-level will change every 2 minutes depending on performance (increase if performance is >85% and decrease if <75%). Sessions consist of 10 ~2 minute blocks, each with n-level as determined by the adaptive procedure and with user paced breaks between blocks.

Improve visual timing and sensitivity in the dorsal stream. The mTBI subject will sit 57 cm in front of a computer monitor. During the presentation, the bars in the 'fish-shaped' window in the center of the screen formed by a sinusoidal grating, move left or right very briefly. When the screen goes blank, the subject reports which way the center pattern moved by pushing the left or right arrow key. A brief tone is presented after incorrect responses. The program adaptively changes the contrast of the test pattern so that the subject detects motion at lowest contrast possible. A sequence of patterns in each training cycle that are designed to optimally activate magnocellular neurons are shown to the subject. There are two programs, the first measures the contrast needed to see one direction of movement, and the second program measures the contrast needed to see two directions of movement, requiring memory.

The sham treatment will be Orientation Discrimination training that is identical to PATH training except instead of low contrast sinewave gratings moving left or right, 100% contrast stationary test and background sinewave gratings are used, both red, green, and black and white gratings. These patterns are randomly oriented left or right, at decreasing tilt angles as the test grating's orientation is identified correctly. These patterns only activate parvocells in ventral pathways (Ungerleider & Mishkin, 1982; Kaplan & Shapley, 1986) instead of activating dorsal pathways, the key components of PATH neurotraining. Therefore, this task does not speed up the brain's visual timing, which is a function of the dorsal stream. For the Orientation Discrimination task, the subject pushes the left arrow key when the test pattern is tilted left and the right arrow key when pattern is tilted right. Otherwise the two training tasks use the same paradigm.

Participants are required to compare each item on a computer screen to the item that they saw n-items back in the sequence. The participant plays a simple game where they control the movement of an astronaut that needs to collect correct gems, avoid incorrect gems, and also obstacles, to succeed at the game. If the participants performs well then they can be advanced to the 2-back, 3-back, 4-back, etc where they make similar matches but to earlier items in the sequence. This gamified task consists of a color n-back with 6 colors (a new color every 3 seconds, requiring subjects respond to targets by tapping the screen and navigating the astronaut to the targets, and avoiding the distractors) with 30% targets, and where the n-level will change every 2 minutes depending on performance (increase if performance is >85% and decrease if <75%). Sessions consist of 10 ~2 minute blocks, each with n-level as determined by the adaptive procedure and with user paced breaks between blocks.

Visual Working Memory (VWM) using Test of Information Processing Skills (TIPS), having two distractor tasks to measure Sequential Processing: the subject must remember a sequence of letters, that are shown one at a time for 2 seconds each, for sequences of from 2 up to 9 letters right after seeing the entire sequence of letters. Short Term that are shown one at a time for 2 seconds each, for sequences of from 2 up to 9 letters right after seeing the entire sequence of letters. Short Term VWM is assessed by recalling the correct sequence of letters after counting from 1 to 10 numbers in sequence, starting at different initial numbers, slowly, and after repeating a short sentence with an animal subject for VWM. Delayed Recall is assessed by remembering all animal names in repeated sentences 3 minutes after finish the VWM tests. The TIPS VWM Standardized Percentile Rank goes from <1% to 99%.

Processing Speed using Wechsler Adult Intelligence Scale (WAIS-4) Processing Speed Index (PSI): requires two subtests: 1) the WAIS Symbol Search subtest which requires subjects to scan a target group (two symbols) and search a group of 5 symbols, indicating whether one of the target symbols appears in the search group, and 2) WAIS Digit Symbol Coding subtest, where the subject fills in boxes below digits with symbols that are paired with them in a key at the top of the page. Both of these subtests are timed for two minutes each. The scaled scores from each subtest are combined to create an overall Processing Speed Index score, that is converted to a standardized percentile. WAIS Processing Speed Index Standardized Percentile Rank goes from < 0.1% to > 99%.

Selective Attention using Delis-Kaplan Executive Function System (DKEFS) Color-Word Interference test measuring time to say color of the printed word that denotes a different color (Stroop test). DKEFS Color-Word Interference Test Standardized Percentile Rank for Inhibition subtest goes from 0.5% to 99.9%.

Cognitive Flexibility using DKEFS Color-Word Interference test measuring time to switch attention between color of printed word (Stroop test) and printed word when surrounded by a rectangle. DKEFS Color-Word Interference Test standardized scores for Standardized Percentile Rank for Inhibition Switching subtest goes from 0.5% to 99.9%.

Auditory Working Memory (AWM) is assessed using the WAIS-4 Working Memory Index which consists of two subtests: 1) the Digit Span subtest, where the subject has to repeat a list of spoken numbers, requiring the subject to remember subsequently more numbers: in the correct order, backwards, and in numerical sequence on three different subtests, and 2) the Letter-Number Sequencing subtest which requires sequencing subsequently more numbers and letters in the correct numerical and alphabetic sequence. Presentation of the numbers and letters are timed for one second each for these working memory tests. The WAIS Working Memory Index Standardized Percentile Rank goes from < 0.1% to > 99.9%.

Using Adult Dyslexia Test, use reading grade level to measure reading proficiency. Evaluates whether a subject's visual processing and phonological processing is markedly, moderately, mildly, or borderline below normal, or is normal or above normal.

Using the total covariance matrix, voxel-wise MEG source magnitude images that cover the whole brain will be obtained for each subject, and each frequency band, following the Fast-VESTAL procedure, measuring time-locked signals during a working memory N-back task to evaluate improvements in brain function in DLPFC (working memory).

Using the total covariance matrix, voxel-wise MEG source magnitude images that cover the whole brain will be obtained for each subject, and each frequency band, following the Fast-VESTAL procedure, measuring time-locked signals during a working memory N-back task to evaluate improvements in brain function in ACC (attention network).

Using the total covariance matrix, voxel-wise MEG source magnitude images that cover the whole brain will be obtained for each subject, and each frequency band, following the Fast-VESTAL procedure, measuring time-locked signals during a working memory N-back task to evaluate improvements in brain function in Precuneus/PCC (attention network).

Using the total covariance matrix, voxel-wise MEG source magnitude images that cover the whole brain will be obtained for each subject, and each frequency band, following the Fast-VESTAL procedure, measuring time-locked signals during a working memory N-back task to evaluate improvements in brain function in V1-MT (visual motion network).

Number of words/minute subject can read 6 words of subsequent text from interesting story using computer-based program.

Using Rivermead Post-Concussion Symptom Questionnaire, evaluates 16 symptoms commonly experienced after a concussion, being rated from 0 (not experienced at all) to 4 (a severe problem). The Rivermead PCS evaluates ocular, vestibular, and mental health symptoms of mTBI subject.

Using ADCS Quality of Life SF-36 questionnaire evaluates subject's general health, activity limitations, emotional health, physical health, social activities, pain, and energy and emotions, having a score from 0-100.

Using Montreal Cognitive Assessment screening test to assess different cognitive domains: attention and concentration, executive functions, memory, language, visuo-constructional skills, conceptual thinking, calculations, and orientation, having a score from 0-30.

Use the standardized blinding procedure and analysis suggested by Liu et al (2016). Joint Estimation of Treatment and Placebo Effects in Clinical Trials with Longitudinal Blinding Assessments. A questionnaire will be administered to all participants, as well as to research assistants who are administering the task. The items will be: three response categories for treatment guess: 'New treatment', 'Placebo (or control)', or 'Don't know (DK)'. The investigators may re-ask those who answered DK initially to make their best guess regarding their assignment. These data can then be used in a joint statistical model to estimate and isolate the relative effect of patient and experimenter effects on treatment efficacy.

A patient with a mTBI who has had a traumatically induced physiologic disruption of brain function will be referred by neurologist Dr. Ahmed who has made a diagnosis of mTBI, which includes one or more of the following (Marshall et al., 2012): any loss of consciousness from 5- 30 min (not longer than 30 min), any loss of memory for events immediately before or after the accident for as much as 24 hours, any alteration of mental state at the time of the accident (e.g. feeling dazed, disoriented, or confused), after 30 minutes, Glasgow Coma Scale of 13-15 (not lower or is considered more severe than a mild TBI), post-traumatic amnesia less than 24 hours, a score of 19-25 on the Montreal Cognitive Assessment (MoCA) screening test, focal neurologic deficits that might/might not be transient, one or multiple concussions. For this study, the following criteria will be utilized: Inclusion Criteria: Diagnosis of mTBI, between the ages of 18 to 55 years, when development and aging are not factors, agrees to complete the study after hearing the time commitment involved, has corrected 20/20 visual acuity, so can do PATH training (dim gray stripes), reads English fluently, so can follow instructions, and can complete the PATH neurotraining task, by pushing the left or right arrow key on the computer. Exclusion Criteria: mTBI occurred less than 3 months earlier, post-traumatic amnesia longer than 24 hours, diagnosis of epilepsy or seizure disorder in last 12 months, diagnosis of major depressive disorder or severe anxiety, answers 'Yes' to any of the questions on the Columbia Suicide Severity Rating Scale, had a stroke or metabolic derangements causing cognitive impairments, ie. alcohol or substance abuse, And for those chosen to undergo MEG exams: has extensive metal dental hardware (e.g., braces and large metal dentures; fillings are acceptable) or other metal objects in head, neck, or face areas that cause artifacts in MEG data, and are not removable during pre-processing, and has claustrophobia since MEG scanner is in small enclosed space.

After completion of the study, deidentified data will be made available to qualified outside investigators following NIH Data Sharing guidelines; none of the data will be considered proprietary. Most of the data for this study will be collected through standardized tests of attention, processing speed, and working memory to be administered by the staff on this project