Development of Point of Care Testing Device for NIRS Based Online Imaging During Electric Brain Stimulation

Transcription

1 Development of Point of Care Testing Device for NIRS Based Online Imaging During Electric Brain Stimulation Thesis submitted in partial fulfillment of the requirements for the degree of MS by Research in Electronics and Communication Engineering by UTKARSH JINDAL International Institute of Information Technology Hyderabad , INDIA August 2016

2 Copyright c Utkarsh Jindal, 2016 All Rights Reserved

3 International Institute of Information Technology Hyderabad, India CERTIFICATE It is certified that the work contained in this thesis, titled Development of Point of Care Testing Device for NIRS Based Online Imaging During Electric Brain Stimulation by Utkarsh Jindal, has been carried out under my supervision and is not submitted elsewhere for a degree. Date Adviser: Dr. Shubhajit Roy Chowdhury Date Co-adviser: Dr. Rahul Shrestha

4 To my Family

5 Acknowledgments I would like to take this opportunity to acknowledge and appreciate the efforts of the people who have helped me during my research and my entire journey through IIIT-H. I wish to express my sincere thanks to Dr. Shubhajit Roy Chowdhury for providing me the opportunity to work with him towards the completion of the project described in this thesis. His guidance, support and immense knowledge always motivated me during the three years I was a part of this project. I also appreciate the efforts of Dr. Rahul Shreshta for being a part this research work. The research atmosphere of IIIT-H also need to be acknowledged as without it I would not have had a chance to explore different research areas. I also wish to express my deep gratitude to Dr. Anirban Dutta, INRIA starting research scientist, DEMAR team for constantly guiding and supervising me in research areas which were new to me. He was always available for all the countless queries I had. I also want to acknowledge Dr. Abhijit Das, Institute of Neurosciences- Kolkata. Without his clinical expertise this project would not have been possible. My thanks and appreciations also go to Ms. Mehak Sood, who willingly helped me out with her abilities throughout the project. The project consumed hard work, lot of research and dedication, and would not have been possible without the moral support of all my friends. I attribute the completion of this research study to the blessings, moral support, love and affection of my parents and my sister. They have always been a major source of motivation and strength.finally, I would like to express my recondite thanks to all those who have rendered their much needed services in the realization of this work. v

6 Abstract In this thesis we present the development of a low cost point of care testing(poct) device for monitoring cerebral oxygenation using near infrared spectroscopy (NIRS). The changes in optical measurements of tissue oxy-(hbo2) and deoxy-(hb) hemoglobin concentration can be monitored using NIRS and these changes can be coupled with anodal transcranial direct current stimulation (tdcs) to find impairments in the cerebral micro-vessels functionality, where impairments in cerebral microvessels functionality may lead to impairments in the cerebrovascular reactivity (CVR). In fact severely reduced CVR predicts the chances of transient ischemic attack and ipsilateral stroke. The feasibility studies on healthy subjects and stroke survivors showed detectable changes in the NIRS responses to anodal tdcs of current density A/m 2. A clinical studyhas been conducted on 14 stroke survivors using our NIRS data acquisition (DAQ) device and our anodal tdcs device. This study revealed that the lesioned hemisphere with impaired circulation showed significantly (p <0.01) less change in HbO2 concentration than the non-lesioned side in response to anodal tdcs. The NIRS sensors used in the study are not disposable as opposed to the present state of the art technology. Our low cost NIRS-DAQ device has been further integrated with the commercially used CE-approved SomaSensor (Covidien, USA). This was done under a clinical study in human subjects of a CE-marked device (being used within indication) to evaluate additional parameters. Here, Medical Devices Directive covers a diverse range of devices that are classified into one of four risk-based Classifications. Our SomaSensor based low-cost NIRS data acquisition device to monitor hemodynamics during noninvasive brain stimulation (NIBS) should be classified as Class IIa - Medium risk, often short-term non-invasive device. The reusability of the SomaSensor was also tested as the commercially used SomaSensors are labelled as disposable. A large chunk of the complete cost of the NIRS setup comes from the procurement of the CE-approved SomaSensors. By proving the reusability of the SomaSensors the complete cost of the NIRS setup was also reduced. The performance of the SomaSensor interfaced with our low cost NIRS-DAQ device was also tested with the help of a dual layer dynamic phantom filled human blood and intralipid. NIRS and anodal tdcs have been further combined with electroencephalography (EEG) to develop a NIRS-EEG/tDCS joint-imaging method. This was done to investigate local post-tdcs alterations from baseline at the site of anodal tdcs using NIRS-EEG/tDCS joint-imaging. In a separate study the changes in local post-tdcs alterations in motor evoked potentials (MEP)-measure of corticospinal excitability were investigated. We found that post-tdcs changes in the mean regional cerebral oxygen vi

7 vii saturation (rso 2 ) from baseline mostly correlated with the corresponding post-tdcs change in logtransformed mean-power of EEG within 0.5Hz Hz frequency band. Moreover, a decrease in log-transformed mean power of EEG within 0.5Hz Hz frequency band corresponded with an increase in the MEP-measure of corticospinal excitability found in the second study. Therefore, we propose to combine NIRS-EEG/tDCS joint-imaging with corticospinal excitability investigation in a single study to confirm these findings.

8 Contents Chapter Page 1 Introduction Background Underlying Concepts Near Infrared Spectroscopy Transcranial Direct Current Stimulation Scope of Present Work Contribution of Thesis Organization of Thesis Development of Hardware for NIRS NIRS system Development Beer Lambert law and its application Sensor implementation for continuous wave (CW) NIRS Anodal tdcs system Development Clinical Validation of NIRS device Introduction Protocol Results Discussion and Future work Interfacing SomaSensor with NIRS-DAQ Introduction Bench-testing CE-approved Somasensor for clinical investigation Graphs of 730nm LED at different current settings Graphs of 810nm LED at different current settings Feasibility testing the single-use SomaSensor for multiple-use in a clinical investigation Performance testing of SomaSensor interfaced with low-cost DAQ using phantom Results and Discussion Development of NIRS-EEG Composite Imaging Montage Introduction Methods NVC evaluation with NIRS-EEG/tDCS joint-imaging Corticospinal excitability changes to anodal tdcs viii

9 CONTENTS ix 5.3 Results NVC evaluation with NIRS-EEG/tDCS joint-imaging Corticospinal excitability changes to anodal tdcs Discussion Conclusions Summary of Present Work Conclusions Scope of Future work

10 List of Figures Figure Page 1.1 Complete spectral range Concept of using NIRS on human brain Variation of spectral extinction coefficient Design of custom made NIRS probe Basic design of amplifier Fourth order Low-pass filter Block Diagram of NIRS setup Pin connections with 7660 voltage converters Positioning of NIRS probe Point of care NIRS hardware Block diagram of anodal tdcs setup Top panel: NIRS detectors (D), and 770nm and 850nm sources (S). Bottom panel: tdcs scalp sites where red denotes the anodal and yellow denotes the cathodal tdcs electrodes Electric field magnitude in the grey matter due to F3 cathodal and Cz anodal tdcs The scatter plot of HbO2 (i.e., change in HbO2 post tdcs from pre-tdcs value) by group (N=14) - left side stroke, right side stroke SomaSensor based low-cost NIRS data acquisition device to monitor hemodynamics during NIBS Relative radiant intensity of 730nm at 40mA Relative radiant intensity of 730nm at 44mA Relative radiant intensity of 730nm at 51mA Relative radiant intensity of 730nm at 56mA Relative radiant intensity of 730nm at 60mA Relative radiant intensity of 730nm at 65mA Relative radiant intensity of 730nm at 71mA Relative radiant intensity of 730nm at 77mA Relative radiant intensity of 810nm at 40mA Relative radiant intensity of 810nm at 44mA Relative radiant intensity of 810nm at 50mA Relative radiant intensity of 810nm at 55mA Relative radiant intensity of 810nm at 60mA Relative radiant intensity of 810nm at 64mA x

11 LIST OF FIGURES xi 4.16 Relative radiant intensity of 810nm at 70mA Relative radiant intensity of 810nm at 81mA The relative radiant intensity of the used 730nm LED at 56mA current The relative radiant intensity of the unused 730nm LED at 56mA current The relative radiant intensity of the used 810nm LED at 56mA current The relative radiant intensity of the unused 810nm LED at 56mA current Phantom preparation Phantom placed on the stirer Phantom placed inside a dark box to reduce external light exposure The complete experimental setup First trial of the first day of phantom testing Second trial of the first day of phantom testing Third trial of the first day of phantom testing First trial of the second day of phantom testing Second trial of the second day of phantom testing Third trial of the second day of phantom testing NIRS-EEG/tDCS joint-imaging montage Electric field magnitude in the grey matter due to F3 cathodal and Cz anodal tdcs Study protocol Change in contralesional hemisphere Change in lesional hemisphere Corticospinal excitability study

12 List of Tables Table Page 2.1 Cost of components Output of NIRS device. left/right denote the side of brain. Pre/post denote the state of stimulation and the values denote the uncalibrated oxygen saturation levels Subjects (M: male, F: female, MCA: middle cerebral artery) xii

13 Chapter 1 Introduction 1.1 Background Stroke or cerebrovascular accident is caused when an artery carrying blood from heart to an area in the brain bursts or a clot obstructs the blood flow thereby preventing delivery of oxygen and nutrients. Worldwide, 15 million people suffer a stroke every year, where Indian studies have estimated that the incidence rates increase from 27-34/100,000 in the age group to /100,000 in the 75+ age group [1][3]. Moreover, ischemic stroke is the leading cause of morbidity and long term disability across the world, and is among the leading causes of death [4]. It is also the second leading cause of dementia (after Alzheimer s disease). The projected cost of patient care for stroke will reach trillions of dollars over the next five decades. Stroke accounts for nearly 10% of deaths and about 5% of health-care costs [5]. The total direct and indirect cost of cardiovascular disease (CVD) and stroke in the United States for 2010 is estimated to be $315.4 billion [6]. This figure includes health expenditures (direct costs, which include the cost of physicians and other professionals, hospital services, prescribed medications, home health care, and other medical durables) and lost productivity that results from premature mortality (indirect costs). In India, stroke in younger individuals is high (18-32% of all stroke cases) compared with the so called developed countries that presents a greater burden on the quality of life [1], [7]. Stroke drastically affects the economical productivity of that younger age group, adding further to the disease burden [7]. With increasing number of younger person with stroke and an aging population, the human as well as the economic burden of stroke and small vessel disease on health systems is likely to dramatically increase in future which is a hidden epidemic for India [8]. In fact, elderly population in India is predicted to increase to 12% of the total population by 2025 [9] where age is the primary risk factor. Consequently, Indian health ministry s working group noted that India will soon become home to the second largest number of elderly in the world. The challenges are unique with this population in India. Therefore, low-cost innovative methodologies for prognosis and monitoring of stroke and small vessel disease are urgently required in India that avoids transportation of critically ill patients and provides home-based intervention. Cross-sectional studies suggest that impaired cerebral haemodynamics precedes ischemic stroke and 1

14 transient ischaemic attack (TIA) where cerebrovascular reactivity (CVR) indicates the capacity of blood vessels to dilate, and is an important marker for brain vascular reserve [10]. CVR provides an useful addition to the traditional baseline blood flow measurement where severely reduced CVR predicts the risk of ipsilateral stroke and TIA [10]. Also, neural activity is closely related, spatially and temporally, to cerebral blood flow (CBF) that supplies glucose via neurovascular coupling. However, neurological symptoms may not be visible until CBF drops to 40% of normal, and prolonged periods of cerebral ischemia increase the likelihood that neurological symptoms will become irreversible as CBF further drops to 20% of normal [11], [12]. The goal of most medical interventions in acute ischemic stroke care is to maximize blood perfusion in the lesioned region and surrounding ischemic penumbra by increasing local and/or global CBF. Intervention strategies to increase CBF are based on the premise that normal auto-regulation of CBF is disrupted and perfusions of the infarcted and penumbral regions are dependent on perfusion pressure [13], [14]. Here, the ability to directly or indirectly measure CBF, concentrations of oxy- and deoxyhemoglobin at the point-of-care/bedside would provide new opportunities for monitoring cerebro-vascular autoregulation in early stroke/tia, permitting hemodynamic interventions to be optimized, and potentially administered before the onset of neurological symptoms. Furthermore, because of the high prevalence of stroke and small vessel disease, and the enormous cost of its management, even strategies with modest benefits are worth pursuing to ameliorate impairments in CBF that may lead to neural impairments leading to Mild Cognitive Impairment (MCI) [15]. Moreover, impaired cerebral haemodynamics may lead to vascular dementia (VD), where the pathophysiological links between VD and Alzheimer disease (AD) has been widely discussed [15], [16]. In fact, an association was found between impaired cerebral microvessels functionality (elucidated with CVR) and unfavorable evolution of cognitive function in patients with AD [16]. These impairments in cerebral microvessels functionality can be detected with the help of near infrared spectroscopy (NIRS) coupled with anodal transcranial direct current stimulation (tdcs) of the affected hemisphere. 1.2 Underlying Concepts Near Infrared Spectroscopy All of us are exposed to optical (i.e., visible and near-infrared) radiation from the sun and other sources throughout our lives. Assuming our eyes are shielded from excessive intensity, and our skin is protected from the ultraviolet content of sunlight, we accept this exposure in the knowledge that it is perfectly safe. Unlike x-rays, optical photons are insufficiently energetic to produce ionisation, and unless light is concentrated to such a high degree that it causes burning to the skin, optical radiation offers no significant hazard. The diagnostic potential of optical methods has been widely known since Jobsis [17] first demonstrated that transmittance measurements of near-infrared (NIR) radiation could be used to monitor the degree of oxygenation of certain metabolites. This led to the development and increasingly widespread use of clinical near-infrared spectroscopy (NIRS), which offers a safe, non-invasive means 2

15 of monitoring cerebral function at the bedside without the use of radioisotopes or other contrast agents [18]. Human tissues contain a variety of substances whose absorption spectra at NIR wavelengths are well defined, and which are present in sufficient quantities to contribute significant attenuation to measurements of transmitted light. The concentration of some absorbers, such as water, melanin, and bilirubin, remain virtually constant with time. However, some absorbing compounds, such as oxygenated haemoglobin (HbO2), deoxyhaemoglobin (Hb), and oxidised cytochrome oxidase (CtOx), have concentrations in tissue which are strongly linked to tissue oxygenation and metabolism. Increasingly dominant absorption by water at longer wavelengths limits spectroscopic studies to less than about 1000 nm. The lower limit on wavelength is dictated by the overwhelming absorption of Hb below about 650 nm. However, within the nm window, it is possible with sensitive instrumentation to detect light which has traversed up to 8 cm of tissue [19]. Figure 1.1 Complete spectral range Functional near infrared spectroscopy (fnirs) is based on the optical measurement of changes in tissue oxy-(hbo 2 ), and deoxy-(hb) haemoglobin concentration whereas cerebral oximetry is cerebral oxygenation monitoring via near-infrared spectroscopy (NIRS) [19]. One aspect of this response - the blood oxygen level dependent (BOLD) signal is the basis for functional magnetic resonance imaging (fmri), a brain imaging modality closely related to fnirs in terms of the underlying measurand. The responses measured during fnirs are usually interpreted in terms of changes in (HbO 2 ) and (Hb) concentration a somewhat richer set of variables than those available from basic fmri. A third hemodynamic variable, (Hbt), can be derived as the sum of (HbO 2 ) and (Hb) concentrations, which is considered a good indicator of the variations in the regional cerebral blood volume (CBV) [20]. NIR wavelengths can be selected such that the change in concentration of oxy-hemoglobin (HbO 2 ) and deoxy-hemoglobin (Hb) in the brain tissue can be detected. Such a spectrophotometric determination 3

16 is based on the absorption maximum of oxyhemoglobin and the isosbestic point of oxyhemoglobin and hemoglobin [21]. NIRS instrumentation works on various measuring principles, e.g., continuous wave (CW), frequency domain (FD), and time domain (TD). Absolute concentration measurements may be possible with more expensive TD and FD techniques [22] but exact quantification was not a crucial factor in our application since we only wanted to detect a relative change in (HbO 2 ) and (Hb) rather than to quantify the hemodynamic response in absolute terms. Figure 1.2 Concept of using NIRS on human brain Transcranial Direct Current Stimulation A notable feature of Transcranial Direct Current Stimulation (tdcs) is that it is a non-invasive way of modulating brain activity. Due to very weak currents and the non-invasiveness of tdcs, this technique is suitable for modulation of the cerebral cortex, the most outer part of the brain, which lies closest to the surface electrodes attached to the patients head. This weak dc current applied non-invasively is strong enough to elicit significant effects on cortical activity. These effects cause the oxygen saturation levels to increase at the stimulated area in the healthy case. It has been shown that anodal tdcs enhanced activity and excitability of the excitatory pyramidal neuron at a population level in a non-specific manner, where µ-rhythm desynchronization was found to be generated [23]. Studies suggests that cathodal tdcs decreases the firing rate of neurons thereby downregulating the activity and excitability of the excitatory pyramidal neuron [24]. Cathodal tdcs induces a decrease in regional cerebral blood flow (rcbf) in cortical and subcortical areas [25]. tdcs involves the use of at least two surface electrodes (one anode and one cathode) to deliver a stimulating current to the patient. Electrical current flows in the direction from anode to cathode for an- 4

17 odal tdcs. It induces polarity-specific changes of cortical blood perfusion [26]. A novel method [27] for capturing cerebrovascular reactivity using near-infrared spectroscopy during transcranial direct current stimulation (tdcs) has been investigated. In fact, a significant correlation between tdcs current strength and increase in regional cerebral blood flow (CBF) has been found [28]. It has been postulated that CBF-alterations are causally related to tdcs-induced alterations in cortical excitability via neuro-vascular coupling. Anodal Transcranial direct current stimulation (tdcs) can increase regional cerebral blood flow (rcbf) during stimulation. Therefore, our objective was to investigate neurovascular coupling through development of a near-infrared spectroscopy device which recorded changes in oxyhaemoglobin and deoxyhaemoglobin concentrations during anodal tdcs-induced activation of the cortical region located under the electrode and in-between the light sources and detectors [29]. Thus combining NIRS with tdcs can be an easy and economical setup for use in clinical population at risk for ischemic stroke. 1.3 Scope of Present Work Monitoring of the brain activity plays an important role in prediction of stroke and transient ischemic attack(tia). Near Infrared Spectroscopy(NIRS) has recently been implemented in various applications. It provides similar haemodynamic information comparable to functional magnetic resonance imaging(fmri) but has an advantage of obtaining measures of both deoxy and oxy-hemoglobin simultaneously. There are several motivating factors for researching the potential of NIR based methods. NIRS equipment offers higher temporal resolution as compared to other imaging methods like fmri. This allows for, among other things, being able to model fast oscillatory noise related to normal physiological functions. Also NIRS equipment is relatively less restraining and safer than other methods like fmri. Using neuroimaging methods like NIRS we may be able to objectively quantify the individual brainstate. This can help in developing treatments to facilitate neuro-rehabilitation by giving an insight into inter-subject variability and intra-subject reliability thereby making it easier to use methodologies like transcranial direct current stimulation (tdcs) for neuro-rehabilitation. Also, a neuroimaging method like NIRS to quantify any homeostatic interaction between the task and tdcs is important to determine their relative timing, e.g., as an adjuvant treatment during motor rehabilitation. Moreover, heterogeneously damaged cortical regions in stroke subjects present a challenge due to unpredictable alterations of current flow and unpredictable effects on the cortical targets, which may need online monitoring of tdcs effects with neuroimaging for driving individualized tdcs as an adjuvant treatment during neurorehabilitation. 1.4 Contribution of Thesis There was no low cost point of care testing(poct) device available for monitoring cerebral oxygenation before the developments in this thesis. In this thesis we have successfully developed the said device 5

18 and made sure that the device is also portable. The studies and clinical trails conducted by us using the device have also confirmed that anodal tdcs can perturb local neural and vascular activity which can be used for assessing regional cerebral microvessels functionality. This has significant clinical scope since impaired cerebral haemodynamics precedes stroke and transient ischemic attack(tia). We have also established that the CE-approved sensors can be used with our custom made NIRS- DAQ thereby eliminating the cost of the commercially available DAQ devices. Acquiring the commercial DAQ devices turn out to be a major problem for clinics and small hospitals in developing countries like India.[7] Through our studies and experiments on investigating the corticospinal excitability changes to anodal tdcs elucidated with NIRS-EEG joint imaging we have proved that the innovative technologies for portable NIRS-EEG neuroimaging may be leveraged to objectively quantify the progress (e.g., corticospinal excitability alterations) and dose tdcs intervention as an adjuvant treatment during neurorehabilitation. 1.5 Organization of Thesis Chapter 2 describes basic principles of NIRS and the development of NIRS-DAQ hardware and sensor implementation in detail. A little overview of hardware of anodal tdcs has also been given. Chapter 3 describes the clinical experiments that were done to validate the NIRS hardware and sensors developed in chapter 2. The clinical study described in this chapter proved to be instrumental in driving the rest of the thesis forward. Further chapter 3 also gives insight for the need of replacing the custom made sensors with CE or FCC approved probes. Chapter 4 describes the development of NIRS-EEG composite imaging montage. This montage was used to study the corticospinal excitability changes to anodal tdcs. This proved useful in proving that NIRS-EEG neuroimaging may be leveraged to objectively quantify the progress and dose tdcs intervention as an adjuvant treatment during neurorehabilitation. Further chapters describe Conclusion, Related Publications and Bibliography. 6

19 Chapter 2 Development of Hardware for NIRS 2.1 NIRS system Development Near Infrared Spectroscopy is used for the purpose of neuroimaging using Near Infrared spectrum. The scalp, skull, and surrounding tissues of the brain exhibit an optical window of tissue transparency in the near-infrared (NIR) range allowing interrogatory measurements of neurochemistry at the cortex. NIRS is a noninvasive imaging method involving the quantification of chromophore concentration resolved from the measurement of near Infrared light attenuation. Using NIRS, brain activity is measured through hemodynamic responses associated with neuron behavior. For this, we use a source-detector pair, where the detector measures the changes in intensity of NIR light provided by the source. From the changes in detected NIR light intensities, it is possible to calculate concentrations of absorbers (termed chromophores) in the tissue such as (Hb) and (HbO). The differences in the absorption spectra of (HbO) and (Hb) allow us to measure the relative changes in hemoglobin concentration over time through use of light attenuation at multiple wavelengths. The next section explains the Beer Lambert law which relates the chromophore concentration change in the tissue to the absorption of light by the chromophore Beer Lambert law and its application The relationship between light intensities and chromophore concentration is commonly expressed using the Beer-Lambert law. A = log( I I 0 ) = ɛ.c.d (2.1) Attenuation (A) is the logarithmic ratio of two intensities, the intensity of the incident light (I 0 ) and the transmitted light (I). ɛ is the molar extinction coefficient, as shown in Fig. 2.1, expressed in mm 1 cm 1. C is the concentration of a chromophore and d is the direct path-length of the photon from the emitting to the 7

20 Figure 2.1 Variation of spectral extinction coefficient receiving optode, i.e. the inter-optode or geometrical distance [30]. Considering that there are more chromophores in the tissue, the Beer-Lambert law can be written using that the absorption coefficient of each chromophore as additive [20], i.e. A = [ɛ 1.C 1 + ɛ 2.C ɛ n.c n ].d (2.2) The typical application of the Beer-Lambert law is that A is experimentally measured (at n wavelengths for n chromophores of interest). Assuming that d is the inter-optode distance and ɛ 1...ɛ n has been found previously for each chromophore, the objective is to find C 1...C n [31]. Accurate estimation of chromophore concentration requires (at a minimum) the same number of wavelengths, as there is chromophores in the given tissue. A similar specific extinction coefficient of Hb and HbO 2 in the nearinfrared spectrum makes measuring at a single wavelength difficult and confounds the interpretation of changes in concentration of the chromophores. Simultaneous measurements of two wavelengths can be used to separate the changes for both Hb and HbO 2 [32]. To increase the sensitivity of the estimation, two wavelengths should be chosen in order to distinguish between Hb and HbO 2 molar extinction coefficients, such that HbO 2 has the highest molar extinction coefficient at one wavelength and the lowest at the other. However, if d is not equal to the inter-optode distance, the hemoglobin changes cannot be quantified and comparisons between different subjects or NIRS regions cannot be made. d depends on subjects characteristics, the measured region, and the wavelength of the light. Due to scattering d will increase in an unknown manner. Approximately 80% of the total attenuation of near infrared light in tissue is 8

21 due to scattering, and the remaining 20% to absorption [18]. Scattering is thus the biggest problem when attempting quantitative measurements with NIRS. In a highly scattering medium, photons travel a mean distance that is far greater than d, which has been defined as the differential path-length (DP), i.e., the true optical distance between the optodes. Delpy et al. [22] defined a scaling factor to correct for the path-length; the differential path-length factor (DPF). Therefore, the modified Beer-Lambert law incorporates the additions. A = log( I I 0 ) = ɛ.c.d.dp F + G (2.3) The attenuation and the specific extinction coefficient in the modified Beer-Lambert law are no longer strictly linearly related, the degree of non-linearity is a function of the scattering coefficients. The scattering coefficient of the tissue together with the geometry of the optodes are described in the term G. G is unknown and therefore an absolute calculation of chromophore concentration that cannot be derived from Eq. (2.3). This is a fundamental problem in tissue NIRS, making the determination of absolute concentrations of Hb and HbO 2 problematic. Assuming that G has the same value for all chromophores in the medium, by using a differential equation between two chromophores, G is cleared. As a consequence, only changes in concentration of chromophores can be measured with NIRS [33]. Assuming that DP F and d remain constant during the measuring period, and d and DP F are known, quantitative data on changes in the concentration of chromophores can be derived. A = ɛ.( C).DP F.d (2.4) The changes in the concentration of chromophores can be calculated using a different wavelength for each chromophore, writing a Beer-Lambert equation for each, and solving the simultaneous equations through matrix inversion [22]. The NIRS technique can still be applied as an oxygenation trend monitor if the DPF is unknown, such as the continuous wave NIRS. There are several techniques available that can be used to calculate the DPF, i.e., by time domain [33], frequency domain [31] or derivative spectroscopy [33], which allow for oxygenation changes to be measured. However, without the knowledge of the term G the absolute chromophore concentration cannot be measured Sensor implementation for continuous wave (CW) NIRS For hardware implementation of sensitive equipment, several factors are crucial for its performance and efficiency. As with any physiological measurement, the standardization of sensor placement, coupling, signal preprocessing, and conditioning together with a well-described experimental paradigm are essential elements in obtaining repeatable performance, optimized signal-to-noise ratio (SNR) and hence reduced workload on the DSP component. Measurement methods for NIRS are based around three distinct technologies: 9

22 Continuous wave (CW) Frequency domain Time domain We are using continuous wave (CW) NIRS in this work. This method generally employ either a multiple discrete wavelength source or a filtered white light source, they measure the light attenuation using either a photomultiplier, photodiode or an avalanche photodiode detector [33]. The main advantages of CW NIRS are the sampling rate, the size of the instrument, the weight, the simplicity and the cost, which makes CW NIRS ideal for bedside monitoring. However, CW NIRS has a few disadvantages, including the penetration depth and the difficulty to separate absorption and scattering effects. CW NIRS is therefore only fit to do oxygenation trend monitoring. The basic design of our NIRS probe along with the alignment of the light emitting diode (LED) and photo-detectors are shown in fig S1 and S2 indicate LED and D1 and D2 are photo-detectors. The distance between the source and detector was maintained at 2.5cm. The wavelength of the source LEDs are 850nm and 760 nm. The photo detector that was used is BPW34S with Dimensions (L x W x H in mm) 5.4 x 4.3 x 3.2 and radiant sensitive area is 7.5 mm 2. The spectral bandwidth range was between nm. The angle of half sensitivity was 65 degrees. Figure 2.2 Design of custom made NIRS probe The low-cost two channels NIRS hardware was implemented that captured the hemodynamic changes in the frontal cortex of the brain due to non-invasive brain stimulation as a measure of cerebrovascular reserve. The signal obtained at two channels (2 detectors) was high pass filtered (f c = 0.8Hz) (using passive high pass filter) and amplified by an instrumentation amplifier (INA131) (see fig 2.3) followed by low pass filtering with a fourth order Butterworth low pass filter (fc = 8.86Hz) (see fig 2.4). The INA131 is a low cost, general purpose low gain, G = 100, instrumentation amplifier offering excellent accuracy. The INA131 is laser trimmed to achieve very low offset voltage (50µV max), drift (0.25µV/ C max), and high CMR (110dB min). For implementing the Butterworth low pass filter, high precision dual operational amplifier IC OPA2277 was used. They offer improved noise, wider output voltage swing and are twice as fast with half the quiescent current. Features include ultralow offset 10

23 Figure 2.3 Basic design of amplifier voltage and drift, low bias current, high common-mode rejection, and high power supply rejection. The block diagram of the setup is illustrated in fig 2.5. Figure 2.4 Fourth order Low-pass filter The analog signal obtained at the output of the low pass filter was converted to digital values using the 10-bit ADC present in Arduino Uno. The digital values were ported to the computer via USB with Arduino Uno serial USART port for processing in MATLAB. The sampling rate for ADC conversion was kept at 80Hz. The Arduino was powered from USB port which was connected to the laptop. The 5V power output of the Arduino was used to power the entire circuit except the LEDs. The LEDs were powered using the digital output pins of the Arduino. The digital output pins of the Arduino were programmed for switching the LEDs ON/OFF systematically. The negative power (-5V) required by OPA2277, INA131 11

24 Figure 2.5 Block Diagram of NIRS setup and photodetectors was generated by using Interstils ICL7660 CMOS voltage converter (see fig. 2.6). For the experimental bench-testing, the NIRS probes were first placed at the FP1/FP2 position (10-20 international system of scalp sites) but after certain trials it was decided that the F3/F4 position see fig. 2.7 give better results as these regions offer better connectivity to the middle cerebral artery(mca). NIRS data consisted of a series of time-dependent voltage signals measured between individual light source and detector positions on a probe. The concentrations of oxygenated, deoxygenated and total hemoglobin were calculated for each source-detector pair. Differential oxy hemoglobin and de-oxy hemoglobin concentrations were then estimated using modified Beer Lamberts Law with the custom software written in MATLAB (The MathWorks, Inc., USA). 12

25 Figure 2.6 Pin connections with 7660 voltage converters Figure 2.7 Positioning of NIRS probe The cost of the major components used in the POCT NIRS hardware are given below in Table 2.1 The sensors attached with the data acquisition circuit in fig 2.8 were replaced with the commercially available CE-approved SomaSensors (Covidien, USA). 13

26 Component Price (in USD) OPA INA AD ICL Arduino Uno 30 Somasensor 300 Table 2.1 Cost of components Figure 2.8 Point of care NIRS hardware 2.2 Anodal tdcs system Development The Anodal tdcs hardware that is developed is capable of providing a constant current ranging from 0.25 ma to 2.25 ma. The current is controlled using a three terminal adjustable current source (lm334) with an analog potentiometer. An auto linear ramp up and ramp down of 30 seconds is applied at the beginning and end of stimulation respectively. The time for which the stimulation lasts can be easily programed from a computer and 4 time values can also be stored on the hardware itself. The ramp up 14

27 and ramp down has been achieved by using a digital potentiometer which is programmed to change its value continuously during the ramp up and ramp down period. The device allows adjustment of output current in fine increments (i.e. 0.1 or 0.25 ma steps) for the protection and comfort of the patient. The block diagram of circuit is shown in fig In case of abnormal skin sensation due to tdcs, the device is provided with a knob with which the intensity of current can be manually reduced. Figure 2.9 Block diagram of anodal tdcs setup Due to ease of availability and economic factors simple sponge electrodes soaked with normal solution (0.9% sodium chloride solution) have been used as anode and cathode [34]. The area of these electrodes is kept to be 5cm 5cm. A saline solution is used to minimize the electrical impedance of the electrode-skin interface and reduce voltage demands for stimulation. Moreover, it removes the possibility of any electrochemical reactions at the electrode-skin interface thereby eliminating the chances of skin lesion. These electrodes were later replaced by commercially available rubber electrodes. 15

28 Chapter 3 Clinical Validation of NIRS device 3.1 Introduction 14 patients with established and subacute ischemic stroke (<1 month) restricted to a single hemisphere (10 males and 4 females between age 42 to 73) were recruited for clinical validation of the developed NIRS device. Subjects were recruited after ethics approval and experiments were conducted at the Institute of Neurosciences Kolkata, India after taking informed consent, conforming to the Declaration of Helsinki.[35] 3.2 Protocol Participants were seated in a quiet room and their eyes were open and fixed on a point on the wall in front of them during the entire experiment. The total testing time was roughly 14 minutes. Anodal tdcs with the anode positioned at Cz (international system of scalp sites) and the cathode over F3 (F4 when monitoring the right side) was conducted with a current density maintained at 0.526A/m 2 (see fig 3.1). This F3 (F4 when monitoring the right side) cathodal and Cz anodal tdcs montage was selected based on computational modeling [34] (see fig 3.1) in order to target primarily the outer convex brain territory (superficial divisions) of the middle cerebral artery (MCA). MCA is one of the three major paired arteries that supply blood to the cerebrum and is the most common site of stroke. Figure 3.2 shows the electric field magnitude in the grey matter due to F3-Cz tdcs montage. The low-cost continuous wave NIRS hardware consisted of four detection channels (D1, D2, D3, D4) that were placed near either F3 (for F3-Cz tdcs montage) or F4 (for F4-Cz tdcs montage) site during two separate trials. In our study, the subjects had ischemic stroke (<1 month) restricted to a single hemisphere, the montages (F3-Cz or F4- Cz) were selected to target the ipsilesional and contralesional hemisphere in separate sessions, with the order randomized across subjects. Here, NIRS was performed for two minutes on the lesioned hemisphere followed by tdcs for three minutes with a montage targeting the lesioned hemisphere. After tdcs that targeted mostly the lesioned hemisphere, NIRS was performed again on 16

29 Figure 3.1 Top panel: NIRS detectors (D), and 770nm and 850nm sources (S). Bottom panel: tdcs scalp sites where red denotes the anodal and yellow denotes the cathodal tdcs electrodes. the lesioned hemisphere to measure any change due to tdcs. The change in (HbO 2 ) post-tdcs from pre-tdcs baseline were compared using paired-sample t-test. HbO 2 = HbO 2post tdcs HbO 2pre tdcs HbO 2post tdcs + HbO 2pre tdcs The same process was repeated with the contralesional hemisphere to find differences (lesional versus contralesional) in the NIRS response since ischemic stroke was restricted to a single hemisphere. 17

30 Figure 3.2 Electric field magnitude in the grey matter due to F3 cathodal and Cz anodal tdcs 3.3 Results The affected hemisphere of the brain with impaired circulation showed significantly less increase in cerebral oxygen saturation as compared to non-lesioned side in response to anodal tdcs. There was significant change in cerebral oxygen saturation post-tdcs from pre-tdcs baseline in the non-lesioned side (3.43 ± 0.86) but not in the lesioned side (0.26 ± 0.28), p<0.01 [36]. Figure 3.3 shows the scatter plot of HbO 2 grouped by left hemispheric (side) stroke and right hemispheric (side) stroke. Table 3.3 shows the output of the NIRS device on the subjects. With the low-cost NIRS-tDCS hardware, it was possible to conduct transcranial cerebral oximetry, i.e., a non-invasive point-of-care technique to measure changes in regional cerebral oxygen saturation following tdcs [36]. NIRS monitored oxygen saturation of hemoglobin in the entire tissue bed, which includes a mixture of brain tissue, arterial and venous blood. It was assumed that anodal tdcs causes changes in synaptic transmembrane current, resulting in change in CBF via a change in the representative radius of the vasculature, and not in arterial and venous blood pressure difference. Furthermore, it was assumed that the 18

31 SNo Left-post Left-pre Right-post Right-pre Table 3.1 Output of NIRS device. left/right denote the side of brain. Pre/post denote the state of stimulation and the values denote the uncalibrated oxygen saturation levels changes from baseline in the NIRS recording of oxygen saturation during tdcs were solely due to the local hemodynamic effects (mainly venous blood) induced by tdcs and not due to changes in arterial oxygen concentration. Also, we assumed that the spatially resolved spectroscopy technique removed skin blood flow artifacts from the NIRS signal. 19

32 Figure 3.3 The scatter plot of HbO2 (i.e., change in HbO2 post tdcs from pre-tdcs value) by group (N=14) - left side stroke, right side stroke. 3.4 Discussion and Future work Our preliminary studies showed the feasibility of identifying the lesioned hemisphere (see fig 3.3) in subacute stroke with the low-cost NIRS-tDCS hardware [36]. A significant change in HbO 2 post-tdcs from pre-tdcs baseline was found on the non-lesioned side (3.43 ± 0.86) but not in the lesioned side (0.26 ± 0.28), p<0.01. In healthy conditions, cerebral blood flow (CBF) should increase in the brain regions with neural activity (facilitated with anodal tdcs) via metabolic coupling mechanisms whereas the autoregulation mechanisms should ensure that the blood flow is maintained during changes in the perfusion pressure. Therefore, it is postulated that impaired blood flow in the lesioned hemisphere is responsible for a decrease (see fig 3.3) in cerebral oxygen saturation in response to anodal tdcs in the lesioned side when compared to that in the non-lesioned side in most stroke survivors. However, it cannot be excluded in certain stroke survivors that the lesioned hemisphere did not respond with an increase in neural activity due to anodal tdcs which will give a similar NIRS response. Specifically, lack of change in neural activity will result in no change in the cerebral metabolic rate of oxygen, CMRO2 (i.e., oxygen consumption) [37], that is given by the difference of oxygen flowing into and out of the tissue. To avoid this detection issue in our subsequent studies, the effects of tdcs on neural activity was elucidated with simultaneous EEG, which provided an independent measure to supplement NIRS recordings. Based on prior work [38] that found an association of individual resting state EEG 20

33 alpha frequency and cerebral blood flow, we hypothesized that if the changes in EEG during tdcs are correlated with the changes in NIRS then it may enable us to measure the state of the neurovascular coupling (NVC) [39]. Also, the custom made sensors used with the NIRS-DAQ device should be replaced by a commercially available CE-approved alternative. This was done by replacing them with SomaSensors (Covidien, USA). The following chapter deals with the interface of somasensors with our NIRS-DAQ, its reusability and its performance testing. 21

34 Chapter 4 Interfacing SomaSensor with NIRS-DAQ 4.1 Introduction For further research it was necessary that the NIRS probe that we use with our NIRS-DAQ are CE or FCC approved. Therefore we explored the options of commercially used sensors and selected INVOS SomaSensors due to ease of availability. INVOS system adult sensors (SomaSensors) are capable of noninvasively monitoring site-specific adequacy of perfusion in the brain or body tissue directly beneath its sensors. This enables us to obtain real-time data on regional oxygen saturation rso 2, which can detect site-specific ischemic complications even when systemic parameters or lab tests are within normal limits.[40],[41],[42],[43] SomaSensors are applied to the skins surface. The sensors are user and patient friendly, making monitoring of ischemic threats to the brain and body safe and easy, while supplying the information the care team needs to assess and intervene accordingly. Just like our custom made sensors, these sensors also have a dual wavelength design. One of the challenges of presented in this chapter is determining the wavelengths of the LED s used in somasensors. Although these sensors are CE-approved, they are labelled as disposable; designed for single use. This poses a huge problem since the goal is to develop a low-cost point of care device. Therefore the reusability of the somasensors has also been tested in section Bench-testing CE-approved Somasensor for clinical investigation Clinical investigation is defined as a Systematic investigation in one or more human subjects, undertaken to assess the safety or performance of a medical device (ISO 14155). The low-cost NIRS sensor developed under INRIA-DST project funding needed a performance evaluation that is defined as Investigation of an in-vitro diagnostic medical device intended to validate the performance claims under the anticipated conditions of use (EN 13612:2002). However, the goal of this study was to prepare the low-cost NIRS data acquisition (DAQ) device to monitor haemodynamics during non-invasive brain stimulation (NIBS) using CE-approved SomaSensor (Covidien, USA) under a clinical study in human 22

35 subjects of a CE-marked device (being used within indication) to evaluate additional parameters. Here, Medical Devices Directive covers a diverse range of devices that are classified into one of four risk-based classifications. Our SomaSensor based low-cost NIRS data acquisition device to monitor hemodynamics during NIBS should be classified as Class IIa - Medium risk, often short-term non-invasive device. These clinical tests were run at the research unit of Dr. Emmanuelle Le Bars et Dr. Nicolas Menjot de Champfleur, Plateforme d Imagerie Fonctionnelle Humaine, Centre Hospitalier Rgional Universitaire (CHRU) de Montpellier, France. Figure 4.1 SomaSensor based low-cost NIRS data acquisition device to monitor hemodynamics during NIBS To successfully interface the SomaSensor (Covidien, USA) with the low-cost DAQ (presented in Chapter 2), the maximum current that can go through the LEDs in the SomaSensor had to be found out experimentally. This was done by varying the current through each LED in the SomaSensor and then looking at the radiant intensity changes along with the changes in the temperature of the SomaSensor. For this, we used a spectrometer Princeton Instrument SP2150 with a CCD Princeton device Pixis Excelon 400B. We also used a diffraction grating with 600 grooves/mm blazed at 500nm and the exposition time was set at 3µs. After studying the graphs obtained by this experiment it was found out that the peak current for both the LED s was 55mA as increasing the current further had no changes on the radiant intensity of the LED s. It was also proved that there are no secondary wavelengths in the LED s as there were no secondary peaks seen in the graphs of wavelength versus radiant intensity. 23

36 4.2.1 Graphs of 730nm LED at different current settings Figure 4.2 Relative radiant intensity of 730nm at 40mA Figure 4.3 Relative radiant intensity of 730nm at 44mA 24

37 Figure 4.4 Relative radiant intensity of 730nm at 51mA Figure 4.5 Relative radiant intensity of 730nm at 56mA 25

38 Figure 4.6 Relative radiant intensity of 730nm at 60mA Figure 4.7 Relative radiant intensity of 730nm at 65mA 26

39 Figure 4.8 Relative radiant intensity of 730nm at 71mA Figure 4.9 Relative radiant intensity of 730nm at 77mA 27

40 4.2.2 Graphs of 810nm LED at different current settings Figure 4.10 Relative radiant intensity of 810nm at 40mA Figure 4.11 Relative radiant intensity of 810nm at 44mA 28

41 Figure 4.12 Relative radiant intensity of 810nm at 50mA Figure 4.13 Relative radiant intensity of 810nm at 55mA 29

42 Figure 4.14 Relative radiant intensity of 810nm at 60mA Figure 4.15 Relative radiant intensity of 810nm at 64mA 30

43 Figure 4.16 Relative radiant intensity of 810nm at 70mA Figure 4.17 Relative radiant intensity of 810nm at 81mA 31

44 4.3 Feasibility testing the single-use SomaSensor for multiple-use in a clinical investigation To successfully interface the SomaSensor (Covidien, USA) with the low-cost DAQ (presented in Chapter 2) for multiple-use in a low-resource clinical setting, tests were run at the research unit of Dr. Emmanuelle Le Bars and Dr. Nicolas Menjot de Champfleur, Plateforme d Imagerie Fonctionnelle Humaine, Centre Hospitalier Rgional Universitaire (CHRU) de Montpellier, France. Two SomaSensors were selected out of which one was unused and unopened before the start of this experiment. The second SomaSensor however had been used for more than 20 times. The radiant intensity and the peak wavelength of both the sensors were compared. The graphs of both the sensors are shown below. We used a spectrometer Princeton Instrument SP2150 with a CCD Princeton Instrument Pixis Excelon 400B. We also used a diffraction grating with 600 grooves/mm blazed at 500nm. The exposition time was set at 3µs. It was found that even after more than 20 times of use, the SomaSensor provided consistent performance when compared to a new SomaSensor. Figure 4.18 The relative radiant intensity of the used 730nm LED at 56mA current 32

45 Figure 4.19 The relative radiant intensity of the unused 730nm LED at 56mA current Figure 4.20 The relative radiant intensity of the used 810nm LED at 56mA current 33

46 Figure 4.21 The relative radiant intensity of the unused 810nm LED at 56mA current 4.4 Performance testing of SomaSensor interfaced with low-cost DAQ using phantom These tests were performed at the laboratory of Prof. Dr. rer.nat. Ulrich G. Hofmann, Peter- Osypka-Professor for Neuroelectronic Systems, University Medical Center Freiburg, Department of Neurosurgery, Germany. For this experiment, 1400mL human blood phantom was prepared. Out of this 1400mL, 28mL was human blood and 14mL was intralipid. It is important to mention here that the blood was donated by a volunteer on the day 1 of the experiment and a part of the same blood sample was used on the day 2 of the experiment. The phantom setup was a dual layer dynamic phantom and the separation between each layer was 1.4cm which has been published by Ms. Almajidy [44]. This separation also included the thickness of the beaker. Fig 4.22 shows the phantom as it is prepared. Both the layers and the separation between them is clearly visible in this figure. The experiments were carried out in a dark room. Fig 4.23 and fig 4.24 show the phantom placed inside a black box which helped in eliminating external light exposure. A magnetic stirrer was used to keep the composition of the phantom uniform. The RPM of the stirrer was maintained at 500. Fig 4.23 shows the stirrer on which the phantom has been placed. For the first 35 minutes, nitrogen was flushed in the phantom to completely remove the oxygen from the phantom. This was followed by bubbling in oxygen at a constant rate for the next 10 minutes. NIRS 34

47 readings were recorded for the entire duration of the experiments. Fig 4.25 shows the complete setup including the phantom and the NIRS device along with data acquisition laptop. The graphs for oxygen saturation obtained after analysis of NIRS data are also shown below. Figure 4.22 Phantom preparation 35

48 Figure 4.23 Phantom placed on the stirer Figure 4.24 Phantom placed inside a dark box to reduce external light exposure 36

49 Figure 4.25 The complete experimental setup The first day of the experiment showed results where the SomaSensor detected the changes in the oxygen saturation in the phantom as shown in fig. 4.26, 4.27 and Figure 4.26 First trial of the first day of phantom testing. 37

50 The initial period (first 25 mins) is when the nitrogen is pumped in the phantom to reduce oxygen saturation and the last part ( 15 mins) is when the oxygen is pumped in the phantom ( 10 mins) to increase oxygen saturation and then the oxygen saturation was maintained for another 5 minutes. Second and third trial of the first day ended early due to a software glitch. Figure 4.27 Second trial of the first day of phantom testing Figure 4.28 Third trial of the first day of phantom testing Unfortunately the second day of the experiment showed no results where the SomaSensor could not detect the changes in the oxygen saturation in the phantom as shown in fig 4.29, 4.30 and

51 We postulate that this was due to deterioration of the phantom as well as low SNR of the SomaSensor interfaced with the low-cost DAQ. The details are discussed in the next section of this chapter. Figure 4.29 First trial of the second day of phantom testing. The initial period (first 25 mins) is when the nitrogen is pumped in the phantom to reduce oxygen saturation and the last part ( 15 mins) is when the oxygen is pumped in the phantom ( 10 mins) to increase oxygen saturation and then the oxygen saturation was maintained for another 5mins. The phantom seems to have deteriorated so much that the SomaSensor could not detect the change in the oxygen saturation. Figure 4.30 Second trial of the second day of phantom testing 39

52 Figure 4.31 Third trial of the second day of phantom testing 4.5 Results and Discussion The low-cost NIRS-DAQ device interfaced with SomaSensors presented in this chapter costs roughly $500 (table 2.1). The optimum current for the SomaSensor was found to be 56mA. Beyond this limit the supplied current resulted in increasing the temperature of the sensor which will damage the sensor over time and will also result in a burn if the sensor is used on a subject. After studying the radiant intensity curves of both the LEDs in both the new and the used SomaSensors we found that the curves for both the new and the used SomaSensor were similar. This led to the conclusion that the SomaSensors can be reused up to 20 times at least. The phantom experiments for performance testing conducted on day 1 showed that the variation in the oxygen saturation to some extent could be captured with the SomaSensor interfaced with low-cost DAQ but the experiments conducted on the day 2 showed no variation in the oxygen saturation levels that could be captured with the SomaSensor interfaced with low-cost DAQ. It is important to mention that all the settings and the protocol was identical on both the days of the experiment. This indicated that the device interfaced with SomaSensors is not sensitive enough to detect the changes in blood phantom that was more than a day old. This problem may be rectified by improving the signal to noise ratio of the SomaSensors after replacong the 10-bit ADC with a 24-bit or a 32-bit ADC. 40

53 Chapter 5 Development of NIRS-EEG Composite Imaging Montage After clinical validation of the NIRS-DAQ device with the custom made sensors in chapter 3 and interfacing the commercially available CE-approved SomaSensors with the NIRS-DAQ in chapter 4 it is time that we study the corticospinal excitability changes to anodal tdcs. In this chapter we will explain these corticospinal excitability changes to anodal tdcs using NIRS-EEG joint imaging. This has been done by running two different clinical studies on the same group of 5 chronic (>6 months) ischemic stroke survivors and comparing their results. The first study evaluates the corticospinal excitability changes to tdcs conducted with single-pulse TMS-based motor evoked potentials (MEP) from resting first dorsal interosseous (FDI) muscle whereas in the second study we use NIRS-EEG joint imaging to observe the changes in log transformed mean EEG power and mean regional cerebral haemoglobin oxygen saturation (rso 2 ). 5.1 Introduction Transcranial direct current stimulation (tdcs) is an electrically based intervention directed at the central nervous system level and has been shown to modulate cortical neural activity and is also a promising tool to facilitate neuroplasticity [45]. However, inter-subject variability and intra-subject reliability limits clinical translation where a recent meta-analyses showed that the treatment effects of tdcs in patients with stroke are rather inconsistent across studies and the evidence for therapeutic efficacy is still uncertain [46]. Here, neuroimaging may be able to objectively quantify the individual brain-state, e.g. cortical excitability alterations, during tdcs which may lend to closed-loop dosing of tdcs [47] as an adjuvant treatment to facilitate neurorehabilitation. Indeed, it is necessary to track the fluctuations in the functional state of the brain where, e.g., alpha-rhythm states may have a significant effect in perceptual learning since more than 60% of the observed inter-subject variability in perceptual learning can be ascribed to ongoing alpha activity [48]. Also, a neuroimaging method to quantify any homeostatic interaction between the task and tdcs is important to determine their relative timing, e.g., as an adjuvant treatment during motor rehabilitation [49]. Moreover, heterogeneously damaged cortical regions in stroke subjects present a challenge due to unpredictable alterations of current flow 41

54 and unpredictable effects on the cortical targets, which may need online monitoring of tdcs effects with neuroimaging for driving individualized tdcs as an adjuvant treatment during neurorehabilitation [50]. Therefore, we hypothesize that neuroimaging-guided brainstate dependent tdcs [51] will lead to its better therapeutic efficacy as an adjuvant treatment in stroke rehabilitation. Non-invasive neuroimaging techniques that have previously been combined with tdcs include electroencephalography (EEG), functional magnetic resonance imaging (fmri), and functional near-infrared spectroscopy (fnirs). fmri relies on an indirect signal, the blood oxygenation level-dependent (BOLD) contrast, which is caused by an increase in oxygen delivery subsequent to increased neuronal activity. fnirs is a relatively recent optical technology for studying brain function that measures several physiological parameters related to cerebral blood flow and oxygenation including measurements of changes in oxygenated (HbO 2 ) and deoxygenated (Hb) hemoglobin. fnirs presents several advantages relative to fmri, such as measurement of concentration changes in both (HbO 2 ) and (Hb) [52], finer temporal resolution, ease of administration and relative insensitivity to movement artefacts. EEG - electrical potential recorded from the scalp as a measure of neural activity [53] - is due to the electric currents from excitable membranes of brain tissue that superimpose at a given location in the extracellular medium and generate the potential. While fmri has become the gold standard for in vivo imaging of the human brain, fnirs and EEG are more convenient and less expensive technology than fmri. Neural activity has been shown to be closely related, spatially and temporally, to cerebral blood flow (CBF) that supplies glucose via neurovascular and metabolic coupling mechanisms [54][55] where excitation versus inhibition acute effects of tdcs on the population kinetics of neural activity can produce a whole spectrum of EEG signals within the oscillatory regime [56]. The effects on the population kinetics depend on the direction of cortical current flow determining the relative influence of acute tdcs on the cellular targets responsible for the modulation of synaptic efficacy, which are primarily the neuron somata and axon terminals [57]. Therefore, not all neural tissue will be equally affected by a given stimulation protocol which may distinctly affect neuronal populations/neuronal compartments [57]. Individual differences in the neural tissue characteristics [53] as well as heterogeneously damaged cortical regions in stroke subjects with respect to the direction of cortical current flow may lead to the variability in response to tdcs [58]. Therefore, we propose concurrent neuroimaging to capture individual factors that determine this variability in response to tdcs. Towards that overarching goal, we developed and showed the feasibility of EEG-NIRS based monitoring of neurovascular coupling (NVC) during tdcs [59]. We present in this chapter experimental results to elucidate the variability in corticospinal excitability changes to tdcs, especially on the lesional hemisphere, based on NIRS- EEG joint-imaging in patients with a previous ischemic stroke of the right or left middle cerebral artery. Furthermore, based on these preliminary results, we propose EEG-NIRS joint-imaging for closed-loop control of tdcs. 42

55 5.2 Methods NVC evaluation with NIRS-EEG/tDCS joint-imaging We conducted simultaneous recording of NIRS and EEG on five chronic (>6 months, see Table 5.1) ischemic stroke survivors after obtaining informed consent from the subjects. The stroke survivors had no contraindications to non-invasive brain stimulation. The experiments are conducted according to the principles expressed in the Declaration of Helsinki [60]. The NIRS-EEG/tDCS joint-imaging protocol is similar to that of our prior study [61]. Participants were seated in a quiet room and their eyes were open and fixed on a point on the wall in front of them during the entire experiment. We place PISTIM (Neuroelectrics, Spain) electrodes over F3 (corresponding to the left hemisphere) and F4 (right hemisphere) of the international EEG system, and SPONSTIM-25 (Neuroelectrics, Spain) electrodes were placed over Cz. This F3 (F4 when monitoring the right hemisphere) anodal and Cz cathodal tdcs montage is selected based on computational modeling (using StimViewer, Neuroelectrics, Spain) [49] in order to target primarily the outer convex brain territory (superficial divisions) of the superficial middle cerebral artery (MCA) territories (see Figures 1a, b). The tdcs at a current density of 0.526A/m 2 is turned ON for 30sec with 10sec ramp-up and ramp-down (see fig 5.3), which is repeated 15 times in random order with 30sec OFF periods in between for the lesional and contralesional hemispheres (ischemic stroke was restricted to a single hemisphere). Eyes-open block-averaged resting-state NIRS oximeter (INVOS Cerebral Oximeter Model 4100, USA) measurements were conducted just above each eyebrow and below the F3/F4 sites using the adult SomaSensor (SAFB-SM, INVOS, USA). The SomaSensor consists of two LED sources (730nm and 810nm) and two photodiode detectors at a distance of 3cm and 4cm so that the short separation NIRS signal can be regressed out from the longer separation NIRS signal in order to diminish the systemic interference [62]. Eyes-open block-averaged resting-state percent change in the mean regional cerebral haemoglobin oxygen saturation (rso 2 ) in the first 10 secs of ON periods, i.e. the onset response [59] [63], relative to the first 10 sec of OFF periods is measured where the baseline is set at the start of the experiment. At the same time we also recorded the eyes-open resting-state EEG (StarStim, Neuroelectrics, Spain) at 500Hz from the nearby electrodes F1, FC3, F5, F2, FC4, F6 (international system) with CMS/DRL electrodes placed in the left mastoid. EEG artifacts related to tdcs are possible due to issues with electrical (e.g. unknown electrode impedance changes during stimulation) and mechanical compatibility (saline from sponges shunting neighboring electrodes) where concurrent recording is possible with an optimized device (StarStim; Neuroelectrics, Spain) and rational experimental design (using PISTIM electrodes; Neuroelectrics, Spain) [64]. We pre-processed the raw EEG using EEGLAB functions (specifically, Artifact Subspace Reconstruction method) [65] and also removed the artefactual epochs following subsequent visual inspection of the data. The log-transformed mean-power of processed EEG within the 0.5Hz-11.25Hz range (selected based on our prior work [59]) from F3, F1, FC3, F5 were averaged for left hemisphere and from the F4, F2, FC4, F6 were averaged for the right hemisphere. The percent change in log-transformed mean-power of EEG within 0.5Hz- 43

56 11.25Hz frequency band was computed for the first 10 secs of ON periods (i.e., the onset response [59] [63]) relative to the first 10 sec of OFF periods. Figure 5.1 NIRS-EEG/tDCS joint-imaging montage Figure 5.2 Electric field magnitude in the grey matter due to F3 cathodal and Cz anodal tdcs 44