Improve recovery from traumatic brain injury by enhancing the function of intact brain regions so people can regain independence and improve their quality of life.

a) Research vision Traumatic brain injury (TBI) is the commonest cause of death and disability in the under 40s (1). Patients who survive are often left with long-lasting disability, and consume large amounts of NHS resources. Cognitive problems, which include impairments of attention, executive function, processing speed and memory, are a major cause of this disability (2). My aim is to improve recovery from traumatic brain injury (TBI) by enhancing cognitive function. Despite substantial investment, the results of trials of acute neuroprotection have been disappointing. A Cochrane review from 2011 concludes: “In the absence of clear evidence of benefit from (acute) neuroprotective drug use, there is an urgent need to explore other potential modulators of late outcome from TBI” (3) An alternative to acute neuroprotection is to attempt to enhance the function of regions that remain intact, but that function inefficiently after TBI. I will pursue this strategy by studying the targeted treatment of: i) dopaminergic dysfunction; and (ii) persistent inflammation. Dopamine has a well-established role in many of the cognitive functions typically impaired by TBI. For example, animal work shows that working memory is sensitive to small changes in dopamine levels, with either too much or too little causing impairment (4). Dopaminergic pathways are vulnerable to damage after TBI (5). Therefore, replacing dopamine in patients who can be shown to be deficient is likely to be an effective strategy. In addition, we have recently shown persistent inflammation after TBI (6). A prolonged and poorly regulated inflammatory response can be damaging (7). Activated microglia can be cytotoxic (8), and their persistent activation might predispose to the development of neurodegenerative conditions such as Alzheimer’s disease in susceptible individuals, or provide the context in which such diseases takes a more aggressive course (7, 9). Hence, treating persistent inflammation might also be beneficial for cognitive function. Integral to my approach is the recognition that TBI cannot be treated as a homogenous disorder. Treatment needs to be targeted, as cognitive impairments after TBI have many potential causes. To address this issue I will build on my previous work and develop diagnostic neuroimaging to: a) allow the causes of persistent cognitive impairment to be accurately identified; and b) allow the effect of treatment to be monitored at the neural level. I will focus on patients with persistent cognitive problems after TBI. This group are frequently disabled by their cognitive difficulties and unable to return to their independent social roles and employment. As the patients are usually young with relatively normal life expectancy this places a massive social and financial burden on family, carers and state. By identifying and treating the cause of these patients cognitive problems I aim to provide improvements that increase independence, improve quality of life and allow a substantial number to return to active employment. (b) Diagnosing traumatic axonal injury (TAI) after traumatic brain injury Aim: To develop an accurate and widely available diagnostic test for TAI. Hypothesis: Diffusion tensor imaging (DTI) can be used to diagnose clinically significant TAI in individual patients. Overview: My recent work shows that diffusion tensor imaging (DTI) can provide information about white matter damage. This allows the severity and consequences of TBI to be more accurately assessed (10-12). As TAI is a key determinant of cognitive function after TBI, and is poorly imaged using standard techniques, DTI is an important diagnostic tool. Most previous DTI work has focused on demonstrating differences between groups. To unlock its full clinical potential, an automated procedure is needed that can be widely used and that provides diagnostic information on individual patients. Currently the best way to achieve this is unclear. This means that although most clinical MRI scanners are capable of acquiring DTI data, the technique is not routinely used in the assessment of TBI. Multivariate pattern classification techniques provide a powerful way of generating accurate diagnostic information [12,13]. When applied to imaging data they allow multiple pieces of information from an image to be integrated to provide the optimum diagnostic decision. Professor Daniel Rueckert at the Biomedical Image Analysis Group (http://biomedic.doc.ic.ac.uk) is a world expert in the use of these techniques for medical image analysis. Together we will use multivariate pattern classification applied to DTI data to develop a technique to diagnose clinically important white matter damage. Our existing work suggests this approach can be used to differentiate patients with memory or executive function impairments. Ongoing work will investigate the use of this approach to target treatment. Methods Stage 1: Developing a predictive model for clinically significant traumatic axonal injury: We will use the unique dataset already collected as part of my ongoing MRC study of civilian TBI outcome and the parallel military BIOSAP study of blast outcome. 125 TBI patients will have been recruited in these studies by the end of 2011(b) Diagnosing traumatic axonal injury (TAI) after traumatic brain injury (Methods cont.) The group consists of 100 civilian patients and 25 soldiers. This cohort, together with 50 neurologically healthy agematched controls, will be used to train a multivariate computer model to automatically diagnose TAI in individual patients (see (10) for DTI method). Detailed clinical outcome data we have collected one year after our initial assessment will allow us to demonstrate whether this information is clinically significant. We will integrate other measures such as injury severity and focal brain injury load to directly assess whether DTI provides unique information about clinical outcome. This approach will also be applied to the diagnosis of blast brain injury. Data from the BIOSAP study will be analysed using this novel approach. This will allow us to directly address whether blast brain injury is associated with a particular pattern of brain injury, which is a key clinical question (13). Stage 2: Developing a usable diagnostic tool: To allow the diagnostic approach to be widely used we will produce a prototype computational algorithm for clinical use. 15 patients will be scanned on three different MR systems with different scanner magnetic field strengths and DTI protocols within IC. This will allow us to assess whether different imaging environments influence the results. The output of the method will be an individual rating of the amount of clinically significant white matter damage and its likely cognitive effects. Once validated, we will link into the common computational platform for medical imaging being developed by Professor Daniel Rueckert in collaboration with the London Imaging Consortium. We will pilot this in an NHS setting in collaboration with Dr Adam Waldman (radiology lead at IC) to provide the foundation for a full evaluation. This will provide the infrastructure necessary to translate our diagnostic approach back into the clinic. Clinical Impact: Accurate diagnosis of traumatic axonal injury will assist the prediction of clinical outcome and provide a mechanism for stratifying current treatment and future clinical trials. (c) Brain network function after TBI – post-traumatic amnesia (PTA) Aim: To provide a diagnostic test of brain network function that can be used acutely after TBI. Hypothesis: Hippocampal connectivity measured with resting fMRI will predict post-traumatic amnesia. Overview: We have previously shown the value of studying brain network function in the chronic phase after TBI (12, 14). In future work we will extend this to the development of a biomarker of brain network function that can be used acutely after TBI. This would allow accurate assessment of the physiological effects of brain injury, and provide a surrogate marker for treatment interventions, allowing shorter periods of initial evaluation and the clarification of mechanisms of action. I will be validated by studying post-traumatic amnesia (PTA). This is a striking cognitive impairment that is present acutely after TBI. It is easily identified through behavioural testing and is a good marker of injury severity (15, 16). Memory function depends upon the interaction of the hippocampus with other brain regions, but until recently it has not been possible to investigate these interactions in vivo. We have recently shown that the structural connections of the hippocampus predict memory impairment in the chronic phase after TBI (10), and that patterns of functional connectivity measured at ‘rest’ with functional MRI predict cognitive impairment (14). Here we will use resting state fMRI in patients during and following PTA to test the hypothesis that hippocampal connectivity predicts the presence and duration of PTA. Using resting state fMRI allows information about the function of memory circuits to be obtained without the need to patients to actively perform a task in the scanner e.g. (17). This dramatically improves how useful the assessment is in a clinical setting, as many patients are unable to perform complex tasks in the scanner. Methods: Patient recruitment: We will recruit 20 patients during PTA in the subacute phase 1-6 weeks after TBI, 20 patients without PTA after TBI, and 20 normal control subjects. The groups will be matched for age and gender. The patient groups will also be matched for severity of head injury and scan timing. Neuroimaging: The two TBI groups will undergo a standard 3T protocol which includes: a) T1 high-resolution structural scan; b) susceptibility weighted imaging (SWI) to assess microbleed load; c) T2 FLAIR imaging; d) DTI; e) resting state fMRI. DTI will assess hippocampal structural connectivity and resting state fMRI functional connectivity. We will use the image analysis methods for resting state fMRI image analysis described in detail previously e.g. (12, 14). The vast majority of patients will have emerged from PTA at 3 months, when they will be scanned again. This will allow a within-subject comparison to identify the neural correlates of PTA. Neuropsychology: Detailed neuropsychological assessment will be performed at both time points. A standardized battery of cognitive tests will be used as described previously e.g. (10), with additional behavioural assessments specific for PTA. Power: We have not previously investigated PTA. However, using a range of effect sizes from our previous work on chronic TBI as a guide (10, 14), we estimate that between 14-20 patients should be sufficient to reliably detect correlations between behavioural measures of outcome and functional MR and DTI measures. Clinical Impact: A reliable biomarker for brain network function would dramatically accelerate the development of interventions aimed a protecting brain function or promoting brain repair acutely after TBI. (d) Diagnosing and treating dopamine dysfunction after TBI Aim: To improve the health of patients by targeting cognitive enhancement with dopaminergic treatment. Hypothesis: Methylphenidate treatment of dopamine deficient patients will improve cognitive function.Overview: In animal models of TBI, reduced dopaminergic function can be observed in the striatum (18, 19). In humans, SPECT tracers are widely available that bind to the dopamine transporter (DAT) and index the integrity of dopaminergic neurons. DAT binding is reduced after TBI, providing direct evidence for damaged dopaminergic pathways (20). In Parkinson’s Disease, the DAT tracer 123I-FP-CIT provides a well validated diagnostic tool that is widely available (21). Hence, 123I-FP-CIT imaging has the potential to be used for clinical assessment of damaged dopamine neurons after TBI, a has the potential to be widely used across the UK. The diagnosis of dopaminergic dysfunction is particularly important because dopaminergic agents have been shown to enhance cognition in several disorders (e.g., ADHD) as well as in the healthy brain. Previous studies have demonstrated that dopaminergic modulation is beneficial for TBI (for review see (5)); however, the evidence base is not consistent (3) and a method for selecting those patients most likely to respond is needed. A major issue limiting pharmacological trials is the substantial heterogeneity in TBI, both in the profile of cognitive impairment and the extent, type and location of the pathology. Furthermore, cognitive function is related in a non-linear way to dopamine levels (22). Therefore, directly diagnosing dopamine deficiency after TBI should provide a rational basis for stratifying patients into subgroups that will respond best to treatment. Methylphenidate (MPH) is a dopaminergic and noradrenergic re-uptake inhibitor. In animals, chronic administration of MPH has been shown to enhance dopamine function in the striatum after TBI (23). In humans, it is an established treatment for attention impairment in ADHD and is a promising candidate for a cognitive enhancer in TBI (5). I will use 123I-FP-CIT imaging to measure dopaminergic function after TBI, and combine this information with advanced MRI assessment of striatal function. This information will be integrated to test whether the treatment effect of MPH on cognition after TBI is dependent on striatal dopaminergic dysfunction. Methods: We will perform a double-blind placebo-controlled cross-over study of MPH in 40 moderate-to-severe TBI patients with persistent impairments in the cognitive domains studied in (10). Similar to previous studies, patients will receive either 0.3 mg/kg of MPH twice a day or placebo, each for one month (24). To minimize the chance of spontaneous recovery during the study patients will be enrolled at least two years after their head injury. Detailed neuropsychometric assessment, as described in (c), will be made on three occasions, before and after each phase of drug administration. Initial drug administration will be in a controlled clinical setting with appropriate clinical monitoring within the John McMichael center in the Hammersmith Hospital. Patient Recruitment will be made from our TBI database. Patients with contraindications to the use of MPH such as cardiovascular problems will be excluded. Neuroimaging: Patients will complete MRI (as in (c)) and 123I-FP-CIT SPECT assessment at study onset. To assess the impact of treatment on brain network function patients will be re-assessed using MRI at two months. Twenty age and gender matched controls will have similar MRI and SPECT scanning on one occasion. To assess and control for potential vascular confounds between groups and between drug and placebo sessions, a breath holding scan will also be included (19). Power analysis of MPH effects on cognitive function suggest an effect size of >0.44 for a range of neuropsychological measures (e.g. (20)). This indicates that group sizes of between 30-40 patients will be adequate to detect an effect of MPH across the whole patient group. A prior study of 123I-FP-CIT imaging in patients after TBI suggests that 2 years after injury) and randomized to receive either minocycline 200mg/day orally or placebo for the first 15 weeks. Treatment with will be switched for the second 15 weeks. Neuropsychological assessment and patient recruitment details are similar to (iii), although a longer duration of treatment will be employed to allow time for any effects of MIN treatment on brain structure to become apparent. Neuroimaging: The MRI assessment protocol described in (c) will be conducted at study initiation and after MIN treatment. We will also use MRI susceptibility mapping to quantify iron load (39), as this may act as a trigger for inflammation. We will study the mechanism of action for MIN in a subset of 15 patients. PET imaging of the TSPO receptor will be performed at the start of the study and following treatment with MIN. A number of TSPO ligands are potentially available, and show improved sensitivity when compared to PK-11195 (used for our previous work). Costs are included for scanning with PBR-28, but a decision about the most appropriate ligand will be made at study initiation. As TSPO binding is likely to be static in the chronic phase post-TBI, PET imaging will not be performed after the placebo treatment phase. MRI controls from (c) and (d) will be used. Blood biomarker assessment: We will also assess a number of markers of neuronal injury and inflammatory response in chronic TBI. These will include polyamines, S100, 14-3-3, IL-6 and TNF-a. Blood samples will be taken at the start of the study and at weeks 15 and 30, and will also be stored for subsequent genetic analysis. Power: The overall MIN study is powered to detect a similar cognitive treatment effect as (d). For the PET component, we have powered the study to detect an effect on TSPO binding in the striatum and thalamus. In animals MIN has been shown to reduce microglial activity by up to 59% (26). In humans, our previous work assessing abnormal microglial activity following TBI suggests that PET-PK11195 is highly sensitive to detecting inflammation in the thalamus and striatum (effect sizes >1). New TSPO ligands have increased sensitivity. Therefore, a conservative estimate is that 11- 14 patients (depending on the brain region considered) will be adequate to detect a treatment effect on TSPO binding. The primary outcome measure will be cognitive function, as assessed by measures of attention and information processing speed. Secondary outcome measures will be: a) microglial activation within the thalamus; b) thalamocortical connectivity assessed using functional MRI and DTI. This design allows assessment of: • Whether MIN improves cognitive functions after TBI. • Whether MIN reduces microglial activation and whether this correlates with cognitive benefit. • The effect of MIN on the functioning of brain networks important for cognition. Clinical Impact: A short course of anti-inflammatory treatment given subacutely or chronically would be a cheap and safe way of improving cognitive function, and potentially protecting against later neurodegeneration. (f) Neuropsychiatric Issues: Neuropsychiatric problems such as depression and anxiety are common and disabling after TBI. In collaboration with Prof. Anne Lingford-Hughes (Consultant Psychiatrist) we will investigate the relationship between our outcome measures and neuropsychiatric problems. These are also key factors in determining outcome after head injury (2), and are often interlinked with cognitive impairment. A common biological basis may be present for cognitive and neuropsychiatric problems and this is an important and neglected area of research (40). Our current work involves an analysis of the relationship between neuropsychiatric symptoms and brain function (e.g. see (14)). The proposed work will involve detailed psychiatric patient assessment using validated questionnaires and semistructured interviews. We will investigate the possible biological basis of neuropsychiatric symptoms post-TBI, as well as assessing the effects of dopaminergic and anti-inflammatory treatment. LaySummary>