Whole brain MRI scans were performed by appropriately qualified personnel at those specified visits (Day 1, 10 and 60 or at early withdrawal [if participant withdrew from study before Day 60 MRI]). Required pulse sequences of diffusion weighted imaging (DWI), T1, and T2 FLAIR was performed to measure lesion volume and to look for the presence of any new acute inflammatory lesions. The investigator or other medically qualified study team member evaluated the Day 10 and 60 scans for any new abnormalities or clinically significant worsening. Digital data for each MRI was sent to a central MRI laboratory for an over-read of the MRI scan and calculation of the lesion volume. Number of participants with change in white matter and demyelination on Day 10 compared to Day 1, Day 60 compared to Day 1 and Day 60 compared to Day 10 were reported.

The clinical chemistry parameters analyzed were albumin, calcium, creatinine, glucose, potassium, sodium, total CO2, alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase, and total bilirubin. Only those parameters for which at least one abnormal value was reported are summarized. The number of participants with abnormal clinical chemistry findings at specified visit were reported.

The clinical chemistry parameters analyzed were white blood cell count, neutrophil count, hemoglobin, platelet count, lymphocytes. Only those parameters for which at least one abnormal value was reported are summarized. The number of participants with abnormal hematology findings at specified visit were reported.

The pharmacokinetic parameters were calculated by standard non-compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. AUC0-t was determined using the linear trapezoidal rule for increasing concentrations and the logarithmic trapezoidal rule for decreasing concentrations. AUC(0-inf) were calculated, where data permit, as the sum of area under the concentration-time curve over the dosing interval from 0 to Day 10 ±1 day (AUC0-10d) and C10d/z, where C10d is the observed plasma concentration at day 10 and z is the terminal phase rate constant calculated after the second dose.

The pharmacokinetic parameters were calculated by standard non- compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. Cmax and Ct were determined directly from the raw concentration-time data.

The pharmacokinetic parameters were calculated by standard non-compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. Tmax and tlast were determined directly from the raw concentration-time data.

The pharmacokinetic parameters were calculated by standard non-compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. The apparent t1/2 obtained as the ratio of natural log (ln)^2/ lambda-Z, where lambda-Z is the terminal phase rate constant estimated by linear regression analysis of the log transformed concentration-time data.

The pharmacokinetic parameters were calculated by standard non-compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. Lambda-Z is the terminal phase rate constant estimated by linear regression analysis of the log transformed concentration-time data.

The pharmacokinetic parameters were calculated by standard non-compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. The clearance was calculated as Dose/ AUC(0-inf).

Gait velocity is an objective, quantitative, reliable, valid and sensitive measure of lower extremity motor recovery in the stroke population. Changes in gait velocity correlates with physical functioning and quality of life. Gait velocity was assessed over a level, indoor 10 meter distance. The time (in seconds) it takes the participant to travel the 10 meter distance was recorded. Participants was asked to walk at their usual or normal pace and may use their normal assistive devices. Baseline assessments were recorded at Day 5 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

BBS is a performance based measure of balance. It is reliable, valid and responsive to change in the stroke population. BBS is a staff-assessed measure that requires the participant to perform 14 activities that evaluate ability to maintain balance. The BBS typically takes 10-15minute to complete. Participants were not allowed to use assistive devices while performing the activities. Each activity was evaluated by direct observation of the participant's performance and was scored on a 5-point ordinal scale (0-4) where a score of 0 represents inability to perform the activity and a score of 4 represents independence in the activity. The minimum total score on the BBS was 0 and maximum was 56. Baseline assessments were recorded at Day 5 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

The FM assessment is a staff-assessed, disease specific, quantitative measure of impairment that is used to assess recovery of sensorimotor function post stroke. The FM is designed to assess the domains of motor function, balance, sensation and joint function. For this study, only the motor function domain was assessed. The motor domain scale takes approximately 30 minutes to complete and evaluates both the upper and lower extremities by direct observation of the participant's performance of 50 items that measure movement, coordination, and reflex action. Each item was scored from 0-2 for a minimum total score of 0 (hemiplegia) and a maximum total score of 100 (normal motor performance). Baseline assessments were recorded at Day 5 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

The Box and Blocks test is an objective, gross manual dexterity test that is reliable and valid in individuals with upper limb impairments. Box and Blocks was a staff-assessed, participant completed test that required the participant to move small wooden blocks from one side of a partitioned box to the other. The score was determined by the number of blocks transferred within a 60 second time period. More number of blocks transferred as compared to Baseline indicated improvement. Both the impaired and normal limbs were tested, starting with the normal limb. The number of blocks transferred were recorded. Baseline assessments were recorded at Day 1 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

Grip strength is an objective measure of arm motor recovery in stroke participants and correlate with functional status and predict recovery. Grip strength was evaluated by a hand grip dynamometer. Three replicate trials was collected for both the impaired and normal hand, starting with the normal hand. Each trial was separated by a resting period of approximately 15-30 seconds. Participants was instructed to squeeze as hard as possible while using a standardized position of grip and the resulting dynamometer reading (in kg) was recorded. The grip strength measures was conducted within approximately 5 minutes. Baseline assessments were recorded at Day 1 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

The mRS was a 6 point scale that measured participant handicap by evaluating limitations in activity and changes in lifestyle. The mRS was staff-assessed and scored from 0=no symptoms at all, 1=no significant disability, 2=slight disability, 3=moderate disability, 4=moderate severe disability, 5=severe disability (severe disability, bedridden, incontinent and requiring constant nursing care and attention). The structured interview was used to administer the mRS. The mRS took approximately 15 minutes to complete when using the structured interview. Number of participants with mRS were reported as per the category of the score.

The NIHSS is a staff-assessed, 15 item, standardized, disease-specific, deficit scale which measures neurological impairment and is used to quantify participant status by measuring the severity of the stroke. The total NIHSS score ranged from 0 (No impairment) to 42 (severe impairment). Approximately 15 minutes were needed to complete the NIHSS. The NIHSS will be collected as part of the eligibility requirements to exclude participants who have a deficit that is either too mild or too severe. Only NIHSS certified study personnel recorded the NIHSS. Baseline assessments were recorded at Day 1 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

The Barthel was a staff-assessed, 10 item activities of daily living index that evaluated feeding, grooming, dressing, excretion (bowels, bladder and toilet skills), bathing, and mobility (transfers, walking, stairs). The total Barthel score ranged from either 0-20 or 0-100 depending on which scoring algorithm was used where 0= unable or dependent and 20 or 100= independent to perform daily activities. For this study, the 100 point scoring algorithm was used. The Barthel takes approximately 5-10 minutes to complete with the participant.

MoCA was an examiner-administered, screening instrument with good validity, reliability, sensitivity and specificity for mild cognitive dysfunction. The MoCA had been studied in stroke participants and was recommended as a tool to monitor and measure cognitive changes post stroke as part of the 2006 National Institute of Neurological Disorders and Stroke - Canadian Stroke Network Vascular Cognitive Impairment Harmonization Standards. The MoCA assesses eight cognitive domains of visuospatial skills, executive function, language, attention, concentration, working memory, memory, and orientation. Participants were asked to complete 14 activities which the examiner scored according to the standardized scoring instructions. While there was no set time limit imposed on a participant. The total MoCA score ranges from 0-30, where 0= worsening and 30 reflects normal cognitive function.

The short form GDS is a measure of depression developed specifically for use in elderly population and is sensitive and valid in the stroke population. The GDS was a participant-completed, 15 item questionnaire where each question referenced how the participant felt over the past week. Each question was answered with either a 'yes' or 'no' response. Of the 15 questions, 10 of them indicate depression when answered 'yes' (questions 2-4, 6, 8-10, 12, 14-15) and 5 indicate depression when answered 'no' (questions 1, 5, 7, 11, 13). Each question received a score of 1 point when the response was indicative of depression. Total score ranged from 0 to 15. the total score ranged from 0-15, where 0 implies no symptoms and higher score implies more severity of symptoms.

TMS is an electrophysiological technique that was used to measure neurologic changes associated with recovery from stroke via alterations in the excitability of the motor system. Motor threshold measures reflect global excitability of the corticospinal pathway, including large pyramidal cells, excitatory/inhibitory interneurons, and spinal motorneurons. Motor threshold was recorded as the lowest stimulus intensity (in percent) eliciting motor evoked potentials (MEPs). Baseline assessments were recorded at Day 1 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post -randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

The serum sample for S100β collected on Day 1 (predose, 1, 6, 12 and 24 hour) and Day 5 and was analyzed for levels. Serum levels of S100β protein were recorded as log transformed values therefore the negative values are reported.

For additional information about this study please refer to the GSK Clinical Study Register

For additional information about this study please refer to the GSK Clinical Study Register

For additional information about this study please refer to the GSK Clinical Study Register

For additional information about this study please refer to the GSK Clinical Study Register

For additional information about this study please refer to the GSK Clinical Study Register

For additional information about this study please refer to the GSK Clinical Study Register

Eligible participants received GSK249320, 5mg/kg of body weight as an IV infusion via a programmable infusion pump. Each participant received two IV doses 9 ±1 days apart, where the first dose was administered 24-72 hours post stroke. Assessments for safety, pharmacokinetics and pharmacodynamics markers were extended to at least 16 weeks.

Eligible participants received GSK249320, 15mg/kg of body weight as an IV infusion via a programmable infusion pump. Each participant received two IV doses 9 ±1 days apart, where the first dose was administered 24-72 hours post stroke. Assessments for safety, pharmacokinetics and pharmacodynamics markers were extended to at least 16 weeks

Total of all reporting groups

Eligible participants received GSK249320, 1 mg/kg of body weight as an IV infusion via a programmable infusion pump. Each participant received two IV doses 9 ±1 days apart, where the first dose was administered 24-72 hours post stroke. Assessments for safety, pharmacokinetics and pharmacodynamics markers were extended to at least 16 weeks.

Eligible participants received GSK249320, 5mg/kg of body weight as an IV infusion via a programmable infusion pump. Each participant received two IV doses 9 ±1 days apart, where the first dose was administered 24-72 hours post stroke. Assessments for safety, pharmacokinetics and pharmacodynamics markers were extended to at least 16 weeks.

Eligible participants received GSK249320, 15mg/kg of body weight as an IV infusion via a programmable infusion pump. Each participant received two IV doses 9 ±1 days apart, where the first dose was administered 24-72 hours post stroke. Assessments for safety, pharmacokinetics and pharmacodynamics markers were extended to at least 16 weeks

The potential clinical importance ranges (low and high) of the vital sign parameters were for systolic blood pressure (SBP) (<85 and >200 millimeter of mercury [mmHg]), diastolic blood pressure (DBP) (<45 and >110 mmHg) and heart rate (HR) (<40 and >110 beats per minute). Only those parameters for which at least one value of potential clinical importance was reported are summarized. The number of participants with potential clinical important vital parameter findings at any visit were reported.

All Subjects population. Only those participants with data available at the indicated time points were analyzed.

Single 12-lead ECGs was obtained. The standard ECG criteria of potential clinical importance were uncorrected QT interval <300 and >600 milliseconds (msec), absolute QTc interval >500 msec, increase from Baseline QTc >60 msec, RR Interval <90 and >2000 msec, PR Interval <110 and >220 msec, QRS Interval <75 and >110 msec. The number of participants with potentially clinically significant ECG abnormality were reported.

All Subjects population. Only those participants with data available at the indicated time points were analyzed.

NCT (electrode placement technique) of sensory and motor function was performed on the unaffected side (i.e., side that is not affected by the stroke) by appropriately qualified personnel at specified visits (Day 5 and 30 and at early withdrawal). Qualified technician performed the testing; however the same neurologist interpreted the NCT data within a single participant. Both upper and lower extremity nerves were tested and the data was recorded. Number of participants with normal and abnormal NCT data were reported.

All Subject population. Only those participants with data available at the indicated time points were analyzed.

Whole brain MRI scans were performed by appropriately qualified personnel at those specified visits (Day 1, 10 and 60 or at early withdrawal [if participant withdrew from study before Day 60 MRI]). Required pulse sequences of diffusion weighted imaging (DWI), T1, and T2 FLAIR was performed to measure lesion volume and to look for the presence of any new acute inflammatory lesions. The investigator or other medically qualified study team member evaluated the Day 10 and 60 scans for any new abnormalities or clinically significant worsening. Digital data for each MRI was sent to a central MRI laboratory for an over-read of the MRI scan and calculation of the lesion volume. Number of participants with change in white matter and demyelination on Day 10 compared to Day 1, Day 60 compared to Day 1 and Day 60 compared to Day 10 were reported.

All Subjects population. Only those participants with data available at the indicated time points were analyzed.

The clinical chemistry parameters analyzed were albumin, calcium, creatinine, glucose, potassium, sodium, total CO2, alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase, and total bilirubin. Only those parameters for which at least one abnormal value was reported are summarized. The number of participants with abnormal clinical chemistry findings at specified visit were reported.

All Subjects population. Data for only those participants available at the indicated time points were collected and analyzed. Data points with null value for participants analyzed indicate data not collected for respective category and treatment arm.

The clinical chemistry parameters analyzed were white blood cell count, neutrophil count, hemoglobin, platelet count, lymphocytes. Only those parameters for which at least one abnormal value was reported are summarized. The number of participants with abnormal hematology findings at specified visit were reported.

All Subjects population. Data for only those participants available at the indicated time points were collected and analyzed. Data points with null value for participants analyzed indicate data not collected for respective category and treatment arm.

Presence of antibodies to GSK249320 were assessed in serum samples of participants using immunoelectro-chemiluminescent assay. Number of participants with positive antibodies to GSK249320 were reported. Only visits where the true positive antibody detection was observed were reported.

All Subjects population. Only those participants with data available at the indicated time points were analyzed.

The pharmacokinetic parameters were calculated by standard non-compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. AUC0-t was determined using the linear trapezoidal rule for increasing concentrations and the logarithmic trapezoidal rule for decreasing concentrations. AUC(0-inf) were calculated, where data permit, as the sum of area under the concentration-time curve over the dosing interval from 0 to Day 10 ±1 day (AUC0-10d) and C10d/z, where C10d is the observed plasma concentration at day 10 and z is the terminal phase rate constant calculated after the second dose.

Pharmacokinetic population comprised of participants from the 'All Subjects' population for whom a pharmacokinetic sample was obtained and analyzed. Only those participants with data available for analysis were analyzed.

The pharmacokinetic parameters were calculated by standard non- compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. Cmax and Ct were determined directly from the raw concentration-time data.

Pharmacokinetic population. Only those participants with data available for analysis were analyzed.

The pharmacokinetic parameters were calculated by standard non-compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. Tmax and tlast were determined directly from the raw concentration-time data.

Pharmacokinetics population. Only those participants with data available for analysis were analyzed.

The pharmacokinetic parameters were calculated by standard non-compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. The apparent t1/2 obtained as the ratio of natural log (ln)^2/ lambda-Z, where lambda-Z is the terminal phase rate constant estimated by linear regression analysis of the log transformed concentration-time data.

Pharmacokinetics population. Only those participants with data available for analysis were analyzed.

The pharmacokinetic parameters were calculated by standard non-compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. Lambda-Z is the terminal phase rate constant estimated by linear regression analysis of the log transformed concentration-time data.

Pharmacokinetics population. Only those participants with data available for analysis were analyzed.

The pharmacokinetic parameters were calculated by standard non-compartmental analysis using Win Nonlin Pro 4.1. All calculations of non-compartmental parameters were based on actual sampling times. The clearance was calculated as Dose/ AUC(0-inf).

Pharmacokinetics population. Only those participants with data available for analysis were analyzed.

Gait velocity is an objective, quantitative, reliable, valid and sensitive measure of lower extremity motor recovery in the stroke population. Changes in gait velocity correlates with physical functioning and quality of life. Gait velocity was assessed over a level, indoor 10 meter distance. The time (in seconds) it takes the participant to travel the 10 meter distance was recorded. Participants was asked to walk at their usual or normal pace and may use their normal assistive devices. Baseline assessments were recorded at Day 5 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

All Subjects population. Only those participants with data available at the indicated time points were analyzed.

BBS is a performance based measure of balance. It is reliable, valid and responsive to change in the stroke population. BBS is a staff-assessed measure that requires the participant to perform 14 activities that evaluate ability to maintain balance. The BBS typically takes 10-15minute to complete. Participants were not allowed to use assistive devices while performing the activities. Each activity was evaluated by direct observation of the participant's performance and was scored on a 5-point ordinal scale (0-4) where a score of 0 represents inability to perform the activity and a score of 4 represents independence in the activity. The minimum total score on the BBS was 0 and maximum was 56. Baseline assessments were recorded at Day 5 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

All Subjects population. Only those participants with data available at the indicated time points were analyzed.

The FM assessment is a staff-assessed, disease specific, quantitative measure of impairment that is used to assess recovery of sensorimotor function post stroke. The FM is designed to assess the domains of motor function, balance, sensation and joint function. For this study, only the motor function domain was assessed. The motor domain scale takes approximately 30 minutes to complete and evaluates both the upper and lower extremities by direct observation of the participant's performance of 50 items that measure movement, coordination, and reflex action. Each item was scored from 0-2 for a minimum total score of 0 (hemiplegia) and a maximum total score of 100 (normal motor performance). Baseline assessments were recorded at Day 5 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

All subjects population. Only those participants with data available at the indicated time points were analyzed.

The Box and Blocks test is an objective, gross manual dexterity test that is reliable and valid in individuals with upper limb impairments. Box and Blocks was a staff-assessed, participant completed test that required the participant to move small wooden blocks from one side of a partitioned box to the other. The score was determined by the number of blocks transferred within a 60 second time period. More number of blocks transferred as compared to Baseline indicated improvement. Both the impaired and normal limbs were tested, starting with the normal limb. The number of blocks transferred were recorded. Baseline assessments were recorded at Day 1 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

All subjects population. Only those participants with data available at the indicated time points were analyzed.

Grip strength is an objective measure of arm motor recovery in stroke participants and correlate with functional status and predict recovery. Grip strength was evaluated by a hand grip dynamometer. Three replicate trials was collected for both the impaired and normal hand, starting with the normal hand. Each trial was separated by a resting period of approximately 15-30 seconds. Participants was instructed to squeeze as hard as possible while using a standardized position of grip and the resulting dynamometer reading (in kg) was recorded. The grip strength measures was conducted within approximately 5 minutes. Baseline assessments were recorded at Day 1 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

All Subjects population. Only those participants with data available at the indicated time points were analyzed.

The mRS was a 6 point scale that measured participant handicap by evaluating limitations in activity and changes in lifestyle. The mRS was staff-assessed and scored from 0=no symptoms at all, 1=no significant disability, 2=slight disability, 3=moderate disability, 4=moderate severe disability, 5=severe disability (severe disability, bedridden, incontinent and requiring constant nursing care and attention). The structured interview was used to administer the mRS. The mRS took approximately 15 minutes to complete when using the structured interview. Number of participants with mRS were reported as per the category of the score.

All subjects population. Only those participants with data available at the indicated time points were analyzed.

The NIHSS is a staff-assessed, 15 item, standardized, disease-specific, deficit scale which measures neurological impairment and is used to quantify participant status by measuring the severity of the stroke. The total NIHSS score ranged from 0 (No impairment) to 42 (severe impairment). Approximately 15 minutes were needed to complete the NIHSS. The NIHSS will be collected as part of the eligibility requirements to exclude participants who have a deficit that is either too mild or too severe. Only NIHSS certified study personnel recorded the NIHSS. Baseline assessments were recorded at Day 1 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post-randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

All Subjects population. Only those participants with data available at the indicated time points were analyzed.

The Barthel was a staff-assessed, 10 item activities of daily living index that evaluated feeding, grooming, dressing, excretion (bowels, bladder and toilet skills), bathing, and mobility (transfers, walking, stairs). The total Barthel score ranged from either 0-20 or 0-100 depending on which scoring algorithm was used where 0= unable or dependent and 20 or 100= independent to perform daily activities. For this study, the 100 point scoring algorithm was used. The Barthel takes approximately 5-10 minutes to complete with the participant.

All subjects population. Only those participants with data available at the indicated time points were analyzed.

MoCA was an examiner-administered, screening instrument with good validity, reliability, sensitivity and specificity for mild cognitive dysfunction. The MoCA had been studied in stroke participants and was recommended as a tool to monitor and measure cognitive changes post stroke as part of the 2006 National Institute of Neurological Disorders and Stroke - Canadian Stroke Network Vascular Cognitive Impairment Harmonization Standards. The MoCA assesses eight cognitive domains of visuospatial skills, executive function, language, attention, concentration, working memory, memory, and orientation. Participants were asked to complete 14 activities which the examiner scored according to the standardized scoring instructions. While there was no set time limit imposed on a participant. The total MoCA score ranges from 0-30, where 0= worsening and 30 reflects normal cognitive function.

All Subjects population. Only those participants with data available at the indicated time points were analyzed.

The short form GDS is a measure of depression developed specifically for use in elderly population and is sensitive and valid in the stroke population. The GDS was a participant-completed, 15 item questionnaire where each question referenced how the participant felt over the past week. Each question was answered with either a 'yes' or 'no' response. Of the 15 questions, 10 of them indicate depression when answered 'yes' (questions 2-4, 6, 8-10, 12, 14-15) and 5 indicate depression when answered 'no' (questions 1, 5, 7, 11, 13). Each question received a score of 1 point when the response was indicative of depression. Total score ranged from 0 to 15. the total score ranged from 0-15, where 0 implies no symptoms and higher score implies more severity of symptoms.

All Subjects population. Only those participants with data available at the indicated time points were analyzed.

TMS is an electrophysiological technique that was used to measure neurologic changes associated with recovery from stroke via alterations in the excitability of the motor system. Motor threshold measures reflect global excitability of the corticospinal pathway, including large pyramidal cells, excitatory/inhibitory interneurons, and spinal motorneurons. Motor threshold was recorded as the lowest stimulus intensity (in percent) eliciting motor evoked potentials (MEPs). Baseline assessments were recorded at Day 1 of the study. The change from Baseline was calculated by subtracting the Baseline values from the individual post -randomization values. If either the Baseline or post-randomization value was missing, the change from Baseline was set to missing as well.

All subjects population. Only those participants with data available at the indicated time points were analyzed.

The serum sample for S100β collected on Day 1 (predose, 1, 6, 12 and 24 hour) and Day 5 and was analyzed for levels. Serum levels of S100β protein were recorded as log transformed values therefore the negative values are reported.

All Subjects population. Only those participants with data available at the indicated time points were analyzed.

Eligible participants received GSK249320, 1 mg/kg of body weight as an IV infusion via a programmable infusion pump. Each participant received two IV doses 9 ±1 days apart, where the first dose was administered 24-72 hours post stroke. Assessments for safety, pharmacokinetics and pharmacodynamics markers were extended to at least 16 weeks.

Eligible participants received GSK249320, 5mg/kg of body weight as an IV infusion via a programmable infusion pump. Each participant received two IV doses 9 ±1 days apart, where the first dose was administered 24-72 hours post stroke. Assessments for safety, pharmacokinetics and pharmacodynamics markers were extended to at least 16 weeks.

Eligible participants received GSK249320, 15mg/kg of body weight as an IV infusion via a programmable infusion pump. Each participant received two IV doses 9 ±1 days apart, where the first dose was administered 24-72 hours post stroke. Assessments for safety, pharmacokinetics and pharmacodynamics markers were extended to at least 16 weeks