PainQx is a medtech company that uses EEG data and proprietary algorithms to objectively measure a patient's pain level. The PainQx system assesses neural activity in the brain's pain matrix and provides scaled pain and side effect biomarkers. These biomarkers directly correlate to a patient's pain state and can be used by pharmaceutical companies to more accurately measure drug efficacy in clinical trials by reducing subjectivity. PainQx provides quantitative reports on a patient's brain activity that identify regions of over/under activation compared to norms, allowing pain levels to be tracked over time and treatment responses to be evaluated objectively.
The document discusses an innovative neuromodulation technique called Scrambler Therapy (ST) for treating Complex Regional Pain Syndrome (CRPS). A study was conducted on 37 patients with CRPS Type I who received 10 ST treatment sessions. Patients reported pain levels before, during, and 6 months after treatment using the Visual Analog Scale (VAS) and Brief Pain Inventory (BPI). Results showed significantly reduced pain scores after ST compared to before. A control group of 42 neuralgia patients undergoing the same ST treatment showed similar pain reductions. The study provides evidence that ST is an effective treatment for reducing chronic neuropathic pain like CRPS.
This document summarizes key findings from neuroimaging studies on pain processing in the brain. A meta-analysis of 122 pain studies found activation in brain regions involved in sensory and affective pain processing, including the thalamus, insula, and anterior cingulate cortex. Studies also show cortical thickness in pain regions correlates with pain modulation abilities and pain catastrophizing traits. Brain plasticity underlies changes from chronic pain, memory of pain, and phantom limb pain.
Three out of ten people suffer from mental disorders but existing medications and therapies are expensive, ineffective, and have side effects. neurogearTM is an effective alternative brain therapy that uses a headset with electromagnets to generate a targeted magnetic field for non-invasive deep brain stimulation. This focuses magnetic stimulation on specific areas of the brain to treat various mental health conditions with greater accuracy than other non-invasive methods like TMS, fewer side effects than invasive therapies, and at a lower cost.
New advances were presented for managing epilepsy, including seizure detection devices, new drug therapies nearing approval, and emerging neuromodulation technologies. Seizure detection devices can help identify unrecognized seizures to reduce injuries, but have issues with false alarms. Promising new therapies include focal brain cooling, silk-based brain implants delivering adenosine, and convection-enhanced drug delivery directly into the brain. Emerging technologies also show potential, such as responsive neurostimulation, optogenetics, MRI-guided laser ablation, and subdural drug infusion devices. These new approaches aim to provide more targeted treatment options for drug-resistant epilepsy.
This document summarizes a research paper that proposes an algorithm to predict epileptic seizures up to 5 minutes in advance using EEG data. The algorithm analyzes EEG recordings to extract seizure prediction characteristics and correlates these with seizure occurrence times. The algorithm could enable preventative therapies by triggering deep brain stimulation to avoid imminent seizures. When tested on 21 patients, the algorithm achieved 81.7% accuracy in predicting seizures up to 5 minutes beforehand, ranging from 80.7-81.5% accuracy for individual brain channels. If validated, this type of advanced seizure prediction could help minimize injuries from sudden seizures.
This document proposes a two-part system to prevent and intervene in nightmares experienced by PTSD patients: 1) A wireless neurofeedback headset would train patients' brainwaves before sleep to prevent nightmares. 2) A contact-free biometric monitoring system would detect nightmares during sleep and use stimuli to wake the patient, reducing nightmare frequency and improving sleep quality. The document reviews evidence that neurofeedback helps manage PTSD symptoms and technology that can monitor biometric data during sleep. It concludes that using these systems could significantly reduce the negative effects of PTSD night terrors through decreased nightmare occurrence.
The document summarizes several research initiatives being conducted by the NMCSD Pain Medicine department, including studies on:
1) Mirror therapy for phantom limb pain, which has shown promising results in reducing pain levels.
2) Intradiscal biacuplasty versus spinal fusion for treating low back pain, with biacuplasty showing reduced pain, improved function and fewer complications compared to fusion.
3) Developing a cricothyroidotomy simulator to enhance procedural training for deployed medical personnel, with initial studies showing improved comfort but moderate ease of use.
Transcranial Magnetic Stimulation ( TMS) for Chronic PainDr. Rafael Higashi
Aula sobre avanço no tratamento da dor crônica com o uso de Estimulação Magnética Transcraniana (EMT) ministrada por Dr. Rafael Higashi, médico neurologista, no departamento de tratamento da dor do Centro Médico da Universidade de Nova York, NYU, EUA.
www.estimulacaoneurologica.com.br
The document discusses an innovative neuromodulation technique called Scrambler Therapy (ST) for treating Complex Regional Pain Syndrome (CRPS). A study was conducted on 37 patients with CRPS Type I who received 10 ST treatment sessions. Patients reported pain levels before, during, and 6 months after treatment using the Visual Analog Scale (VAS) and Brief Pain Inventory (BPI). Results showed significantly reduced pain scores after ST compared to before. A control group of 42 neuralgia patients undergoing the same ST treatment showed similar pain reductions. The study provides evidence that ST is an effective treatment for reducing chronic neuropathic pain like CRPS.
This document summarizes key findings from neuroimaging studies on pain processing in the brain. A meta-analysis of 122 pain studies found activation in brain regions involved in sensory and affective pain processing, including the thalamus, insula, and anterior cingulate cortex. Studies also show cortical thickness in pain regions correlates with pain modulation abilities and pain catastrophizing traits. Brain plasticity underlies changes from chronic pain, memory of pain, and phantom limb pain.
Three out of ten people suffer from mental disorders but existing medications and therapies are expensive, ineffective, and have side effects. neurogearTM is an effective alternative brain therapy that uses a headset with electromagnets to generate a targeted magnetic field for non-invasive deep brain stimulation. This focuses magnetic stimulation on specific areas of the brain to treat various mental health conditions with greater accuracy than other non-invasive methods like TMS, fewer side effects than invasive therapies, and at a lower cost.
New advances were presented for managing epilepsy, including seizure detection devices, new drug therapies nearing approval, and emerging neuromodulation technologies. Seizure detection devices can help identify unrecognized seizures to reduce injuries, but have issues with false alarms. Promising new therapies include focal brain cooling, silk-based brain implants delivering adenosine, and convection-enhanced drug delivery directly into the brain. Emerging technologies also show potential, such as responsive neurostimulation, optogenetics, MRI-guided laser ablation, and subdural drug infusion devices. These new approaches aim to provide more targeted treatment options for drug-resistant epilepsy.
This document summarizes a research paper that proposes an algorithm to predict epileptic seizures up to 5 minutes in advance using EEG data. The algorithm analyzes EEG recordings to extract seizure prediction characteristics and correlates these with seizure occurrence times. The algorithm could enable preventative therapies by triggering deep brain stimulation to avoid imminent seizures. When tested on 21 patients, the algorithm achieved 81.7% accuracy in predicting seizures up to 5 minutes beforehand, ranging from 80.7-81.5% accuracy for individual brain channels. If validated, this type of advanced seizure prediction could help minimize injuries from sudden seizures.
This document proposes a two-part system to prevent and intervene in nightmares experienced by PTSD patients: 1) A wireless neurofeedback headset would train patients' brainwaves before sleep to prevent nightmares. 2) A contact-free biometric monitoring system would detect nightmares during sleep and use stimuli to wake the patient, reducing nightmare frequency and improving sleep quality. The document reviews evidence that neurofeedback helps manage PTSD symptoms and technology that can monitor biometric data during sleep. It concludes that using these systems could significantly reduce the negative effects of PTSD night terrors through decreased nightmare occurrence.
The document summarizes several research initiatives being conducted by the NMCSD Pain Medicine department, including studies on:
1) Mirror therapy for phantom limb pain, which has shown promising results in reducing pain levels.
2) Intradiscal biacuplasty versus spinal fusion for treating low back pain, with biacuplasty showing reduced pain, improved function and fewer complications compared to fusion.
3) Developing a cricothyroidotomy simulator to enhance procedural training for deployed medical personnel, with initial studies showing improved comfort but moderate ease of use.
Transcranial Magnetic Stimulation ( TMS) for Chronic PainDr. Rafael Higashi
Aula sobre avanço no tratamento da dor crônica com o uso de Estimulação Magnética Transcraniana (EMT) ministrada por Dr. Rafael Higashi, médico neurologista, no departamento de tratamento da dor do Centro Médico da Universidade de Nova York, NYU, EUA.
www.estimulacaoneurologica.com.br
This document presents 4 case reports on using medical shockwave therapy to treat complex and neuropathic pain syndromes in the lower extremities. The patients received 3 treatments of low-intensity shockwaves over 3 weeks and experienced reductions in pain levels and improvements in function based on questionnaires. The results provide preliminary evidence that shockwave therapy may help treat neuropathic and complex pain conditions and warrant further investigation.
Abnormal changes in cortical activity in women with migraine bet.docxdaniahendric
The document describes a study that used magnetoencephalography (MEG) to investigate differences in cortical excitability between migraine patients and healthy controls. MEG recordings were obtained from 35 migraine patients experiencing a headache and 35 healthy controls during an auditory-motor task. Analysis of low and high frequency neuromagnetic signals at sensor and source levels found that high frequency signals showed greater heterogeneous cortical activation in migraine patients compared to controls. The degree of heterogeneous cortical activation correlated with headache frequency. This suggests cortical excitability is altered in a spatially heterogeneous and frequency dependent manner in migraine patients.
Spinesense: Advanced Pain Detection System for Spinal CordIRJET Journal
This document discusses an advanced pain detection system called Spinesense that uses machine learning to analyze MRI and galvanic skin response data to detect spinal cord pain. The system was trained on a dataset containing over 8,600 pieces of data with 440 parameters using random forest classifiers. It evaluates the use of explainable AI to demonstrate how AI can help assess pain levels. The goal is to improve automated pain detection and make it more resistant to variations in individual pain sensitivity and intensity. A time series analysis using the FB Prophet method was also conducted on the spinal cord dataset to enable future prediction of pain levels.
A great presentation by Nathaniel Katz, MD, MS (CEO and owner, Analgesic Solutions) to the FDA in 2011 on sources of measurement error in pain clinical trials.
How to measure and improve brain-based outcomes that matter in health careSharpBrains
Pioneers advancing health research, prevention and treatment will help us understand emerging best practices where targeted assessments, monitoring and interventions can transfer into significant healthcare and quality of life outcomes.
-- Chair: Alvaro Fernandez, CEO & Co-Founder of SharpBrains
-- Dr. Madeleine S Goodkind, staff psychologist at New Mexico VA Health Care System
-- Dr. Randy McIntosh, Vice-president of Research and Director of Baycrest’s Rotman Research Institute
-- Chris Berka, CEO and Co-Founder of Advanced Brain Monitoring (ABM)
Presentation @ The 2015 SharpBrains Virtual Summit http://sharpbrains.com/summit-2015/agenda
This document discusses the use of machine learning techniques and brain imaging to develop biomarkers for classifying chronic musculoskeletal pain conditions. It describes how machine learning algorithms can be trained on brain MRI data to separate chronic pain patients from healthy controls with 70-92% accuracy. However, it notes that self-report remains the gold standard for pain assessment. The document outlines several scientific, practical, and ethical considerations regarding the clinical application of pain biomarkers, such as the need for high reliability, generalizability across large samples, and adequate performance based on the context of use. While biomarkers may provide objective measures of chronic pain, the document cautions that their predictive value depends on the prevalence of conditions.
Prediction of Neurological Disorder using Classification ApproachBRNSSPublicationHubI
This document summarizes a research article about predicting neurological disorders using classification approaches. The researchers tested various classification algorithms on brain morphology data to distinguish between healthy individuals and patients with neurological disorders. They found that the KNN-MinMax classifier was the most effective, achieving prediction accuracy rates 6-8% higher than existing methods. The proposed methodology involved feature extraction, selection, and classification of the data using algorithms like KNN, PCA, random forest, and neural networks.
"Ultrasound- or Nerve Stimulation-Guided Wrist Blocks for Carpal Tunnel Relea...Lucie Beylacq
This study compared ultrasound-guided wrist blocks to nerve stimulation-guided wrist blocks for carpal tunnel release. Sixty patients were randomly assigned to receive blocks using either ultrasound or nerve stimulation guidance. The time to perform the blocks and the onset time of sensory block were measured. Ultrasound guidance took less time to perform the median and ulnar nerve blocks but nerve stimulation had a faster onset of sensory block. However, the success rate of achieving a complete sensory block was the same at 93% for both techniques. This study demonstrates that ultrasound guidance is as effective as nerve stimulation for wrist blocks for carpal tunnel release.
This document discusses predictive neurostimulation (PNS), a new technology for treating intractable seizures. PNS uses a device implanted in the brain to identify seizure-related neurological activity 4-12 minutes before onset and suppresses the seizures without any discomfort. Initial data from 10 patients showed 86% accuracy in detecting seizures and the system stopped 100% of detected seizures. PNS offers clinicians a new treatment option for the over 1 million epilepsy patients who do not benefit from drugs, with fewer side effects than existing options like VNS. The US epilepsy market represents a $20 billion opportunity.
This document discusses using biorobotics to detect diseases like Motor Neuron Disease, Parkinson's Disease, and Alzheimer's Disease. It describes the DDX system, a portable device that measures response time, speed, and force using a joystick, buttons, and sensors. The DDX analyzes parameters like reaction time, tremor, and pressure using fuzzy logic to provide intelligent disease detection. It has advantages like portability and remote monitoring capabilities. Future work may include improved hardware, additional diagnostic parameters, and self-learning techniques to enhance the system's performance.
Neural blockade for persistent pain after breast cancer surgery Jason Attaman
1) The review examined evidence for neural blockade as a diagnostic tool or treatment for persistent pain after breast cancer surgery.
2) Only 7 studies with a total of 135 patients were identified that used blocks targeting the stellate ganglion, paravertebral plexus, or intercostal nerves.
3) The quality of evidence from the studies was low and inconclusive about the efficacy of neural blockade for treating persistent pain after breast cancer surgery. More high-quality studies are needed to evaluate this common clinical problem.
This document proposes a third mechanistic descriptor for chronic pain states characterized by altered nociceptive function. The current definitions of nociceptive and neuropathic pain leave some patient groups without a valid descriptor. The authors argue that a third term, such as "nociplastic," is needed to describe conditions like fibromyalgia, CRPS type 1, and chronic low back pain where nociceptive function is altered despite no clear tissue damage or nervous system pathology. They believe this could improve diagnosis and treatment of affected patients by recognizing altered nociception as an important mechanism. However, critics may argue that nociceptive and neuropathic mechanisms are better established than inferred changes in nociceptive processing.
This document summarizes machine learning methods to learn improved EEG biomarkers in clinical trials from neural activity data. It discusses using machine learning techniques like convolutional neural networks to break down complex neural signals into interpretable patterns that may be related to clinical conditions, outcomes, susceptibility, or treatment responses. As a proof of concept, the document applies these methods to EEG data from a clinical trial of autism treatments, showing the machine learning approach can better classify treatment stage compared to traditional analysis methods and help identify potential neural biomarkers. It also discusses challenges in generalizing models to new patients and adapting methods to work with small clinical trial datasets.
Pneumothorax Detection Using Deep Convolutional Neural NetworksMichael Sebetich
Pneumothorax (PTX), more commonly known as “collapsed lung”, is a potentially life-threatening condition in which air enters the cavity between a patient’s lung and chest wall, inhibiting their ability to breathe. If not treated in a timely manner, typically through the insertion of a chest tube, PTX can result in hypoxia (oxygen deprivation), neurological damage, and, in the worst case, death. Diagnosis of PTX is currently a manual process performed by a radiologist via inspection of a patient’s chest x-ray image. In this paper we explore the potential of Deep Convolutional Neural Networks (CNNs) to automatically diagnose PTX in chest x-rays with the hope of reducing diagnosis time of this condition. We also present heat mapping based on network activations as a novel technique to visualize its performance against individually classified images. The NIH ChestX-ray8 dataset, which is labeled and contains over 100,000 anonymized chest x-rays from 30,000 patients, was used to train a Deep CNN. The final trained CNN is comprised of 5 convolutional layers, 4 pooling layers and 3 dropout layers. This network has a prediction accuracy of 78.5% and a ROC of 0.86 on the validation dataset. These results are encouraging and indicate that with further development Deep Learning has the potential to be clinically useful for automated Pneumothorax detection
ZMPCZM016000.11.23 Electrotherapy for pain managementpainezeeman
This document summarizes research on the use of electrotherapy/electrical stimulation for pain management. It discusses two major theories for how electrotherapy relieves pain through gate control and opiate-mediated control. Research studies cited found electrotherapy effective at reducing pain and improving function for chronic musculoskeletal pain, low back pain, and post-operative knee pain. Meta-analyses showed significant decreases in pain from electrical nerve stimulation and reductions in analgesic consumption when using adequate stimulation parameters.
Neurobionics and robotic neurorehabilitationsNeurologyKota
This document discusses neurobionics, robotic neurorehabilitation, and applications of neurobionics. It summarizes key areas including: (1) neurobionics aims to integrate electronics with the nervous system to repair or substitute impaired functions, (2) robotic neurorehabilitation uses robots to assist in rehabilitation processes, and (3) applications of neurobionics include motor interfaces like robotic arms, sensory interfaces like cochlear implants, and treating conditions like epilepsy and Parkinson's disease.
1) A previous study found that repetitive painful stimulation led to decreased pain ratings over 8 days, associated with increased activity in the rACC region.
2) A 1-year follow-up of 10 participants from the original study found pain thresholds had returned to baseline levels. Pain ratings and rACC activity patterns also returned to initial levels.
3) These findings indicate that habituation to pain involves increased activity of the endogenous antinociceptive system, but this is a temporary adaptive process that recedes once painful stimulation ends, though it may last beyond the stimulation period.
This document describes a project aimed at classifying Obstructive Sleep Apnea Syndrome (OSAS) and Major Depressive Disorder (MDD) in patients using machine learning techniques. The project uses a dataset containing neurocognitive measures like EEG signal features and physiologic measures like heart rate and blood pressure. It aims to develop Python routines using QDA classification and RBF neural networks to analyze the data and accurately classify the severity of patients' OSAS, which current medical experts have difficulty doing by manually analyzing multiple data sources. The project is designed in modules and uses software tools like Python, PyQt, and SQLite3 for implementation.
More Related Content
Similar to PainQx -Science Business Development 3.21.16
This document presents 4 case reports on using medical shockwave therapy to treat complex and neuropathic pain syndromes in the lower extremities. The patients received 3 treatments of low-intensity shockwaves over 3 weeks and experienced reductions in pain levels and improvements in function based on questionnaires. The results provide preliminary evidence that shockwave therapy may help treat neuropathic and complex pain conditions and warrant further investigation.
Abnormal changes in cortical activity in women with migraine bet.docxdaniahendric
The document describes a study that used magnetoencephalography (MEG) to investigate differences in cortical excitability between migraine patients and healthy controls. MEG recordings were obtained from 35 migraine patients experiencing a headache and 35 healthy controls during an auditory-motor task. Analysis of low and high frequency neuromagnetic signals at sensor and source levels found that high frequency signals showed greater heterogeneous cortical activation in migraine patients compared to controls. The degree of heterogeneous cortical activation correlated with headache frequency. This suggests cortical excitability is altered in a spatially heterogeneous and frequency dependent manner in migraine patients.
Spinesense: Advanced Pain Detection System for Spinal CordIRJET Journal
This document discusses an advanced pain detection system called Spinesense that uses machine learning to analyze MRI and galvanic skin response data to detect spinal cord pain. The system was trained on a dataset containing over 8,600 pieces of data with 440 parameters using random forest classifiers. It evaluates the use of explainable AI to demonstrate how AI can help assess pain levels. The goal is to improve automated pain detection and make it more resistant to variations in individual pain sensitivity and intensity. A time series analysis using the FB Prophet method was also conducted on the spinal cord dataset to enable future prediction of pain levels.
A great presentation by Nathaniel Katz, MD, MS (CEO and owner, Analgesic Solutions) to the FDA in 2011 on sources of measurement error in pain clinical trials.
How to measure and improve brain-based outcomes that matter in health careSharpBrains
Pioneers advancing health research, prevention and treatment will help us understand emerging best practices where targeted assessments, monitoring and interventions can transfer into significant healthcare and quality of life outcomes.
-- Chair: Alvaro Fernandez, CEO & Co-Founder of SharpBrains
-- Dr. Madeleine S Goodkind, staff psychologist at New Mexico VA Health Care System
-- Dr. Randy McIntosh, Vice-president of Research and Director of Baycrest’s Rotman Research Institute
-- Chris Berka, CEO and Co-Founder of Advanced Brain Monitoring (ABM)
Presentation @ The 2015 SharpBrains Virtual Summit http://sharpbrains.com/summit-2015/agenda
This document discusses the use of machine learning techniques and brain imaging to develop biomarkers for classifying chronic musculoskeletal pain conditions. It describes how machine learning algorithms can be trained on brain MRI data to separate chronic pain patients from healthy controls with 70-92% accuracy. However, it notes that self-report remains the gold standard for pain assessment. The document outlines several scientific, practical, and ethical considerations regarding the clinical application of pain biomarkers, such as the need for high reliability, generalizability across large samples, and adequate performance based on the context of use. While biomarkers may provide objective measures of chronic pain, the document cautions that their predictive value depends on the prevalence of conditions.
Prediction of Neurological Disorder using Classification ApproachBRNSSPublicationHubI
This document summarizes a research article about predicting neurological disorders using classification approaches. The researchers tested various classification algorithms on brain morphology data to distinguish between healthy individuals and patients with neurological disorders. They found that the KNN-MinMax classifier was the most effective, achieving prediction accuracy rates 6-8% higher than existing methods. The proposed methodology involved feature extraction, selection, and classification of the data using algorithms like KNN, PCA, random forest, and neural networks.
"Ultrasound- or Nerve Stimulation-Guided Wrist Blocks for Carpal Tunnel Relea...Lucie Beylacq
This study compared ultrasound-guided wrist blocks to nerve stimulation-guided wrist blocks for carpal tunnel release. Sixty patients were randomly assigned to receive blocks using either ultrasound or nerve stimulation guidance. The time to perform the blocks and the onset time of sensory block were measured. Ultrasound guidance took less time to perform the median and ulnar nerve blocks but nerve stimulation had a faster onset of sensory block. However, the success rate of achieving a complete sensory block was the same at 93% for both techniques. This study demonstrates that ultrasound guidance is as effective as nerve stimulation for wrist blocks for carpal tunnel release.
This document discusses predictive neurostimulation (PNS), a new technology for treating intractable seizures. PNS uses a device implanted in the brain to identify seizure-related neurological activity 4-12 minutes before onset and suppresses the seizures without any discomfort. Initial data from 10 patients showed 86% accuracy in detecting seizures and the system stopped 100% of detected seizures. PNS offers clinicians a new treatment option for the over 1 million epilepsy patients who do not benefit from drugs, with fewer side effects than existing options like VNS. The US epilepsy market represents a $20 billion opportunity.
This document discusses using biorobotics to detect diseases like Motor Neuron Disease, Parkinson's Disease, and Alzheimer's Disease. It describes the DDX system, a portable device that measures response time, speed, and force using a joystick, buttons, and sensors. The DDX analyzes parameters like reaction time, tremor, and pressure using fuzzy logic to provide intelligent disease detection. It has advantages like portability and remote monitoring capabilities. Future work may include improved hardware, additional diagnostic parameters, and self-learning techniques to enhance the system's performance.
Neural blockade for persistent pain after breast cancer surgery Jason Attaman
1) The review examined evidence for neural blockade as a diagnostic tool or treatment for persistent pain after breast cancer surgery.
2) Only 7 studies with a total of 135 patients were identified that used blocks targeting the stellate ganglion, paravertebral plexus, or intercostal nerves.
3) The quality of evidence from the studies was low and inconclusive about the efficacy of neural blockade for treating persistent pain after breast cancer surgery. More high-quality studies are needed to evaluate this common clinical problem.
This document proposes a third mechanistic descriptor for chronic pain states characterized by altered nociceptive function. The current definitions of nociceptive and neuropathic pain leave some patient groups without a valid descriptor. The authors argue that a third term, such as "nociplastic," is needed to describe conditions like fibromyalgia, CRPS type 1, and chronic low back pain where nociceptive function is altered despite no clear tissue damage or nervous system pathology. They believe this could improve diagnosis and treatment of affected patients by recognizing altered nociception as an important mechanism. However, critics may argue that nociceptive and neuropathic mechanisms are better established than inferred changes in nociceptive processing.
This document summarizes machine learning methods to learn improved EEG biomarkers in clinical trials from neural activity data. It discusses using machine learning techniques like convolutional neural networks to break down complex neural signals into interpretable patterns that may be related to clinical conditions, outcomes, susceptibility, or treatment responses. As a proof of concept, the document applies these methods to EEG data from a clinical trial of autism treatments, showing the machine learning approach can better classify treatment stage compared to traditional analysis methods and help identify potential neural biomarkers. It also discusses challenges in generalizing models to new patients and adapting methods to work with small clinical trial datasets.
Pneumothorax Detection Using Deep Convolutional Neural NetworksMichael Sebetich
Pneumothorax (PTX), more commonly known as “collapsed lung”, is a potentially life-threatening condition in which air enters the cavity between a patient’s lung and chest wall, inhibiting their ability to breathe. If not treated in a timely manner, typically through the insertion of a chest tube, PTX can result in hypoxia (oxygen deprivation), neurological damage, and, in the worst case, death. Diagnosis of PTX is currently a manual process performed by a radiologist via inspection of a patient’s chest x-ray image. In this paper we explore the potential of Deep Convolutional Neural Networks (CNNs) to automatically diagnose PTX in chest x-rays with the hope of reducing diagnosis time of this condition. We also present heat mapping based on network activations as a novel technique to visualize its performance against individually classified images. The NIH ChestX-ray8 dataset, which is labeled and contains over 100,000 anonymized chest x-rays from 30,000 patients, was used to train a Deep CNN. The final trained CNN is comprised of 5 convolutional layers, 4 pooling layers and 3 dropout layers. This network has a prediction accuracy of 78.5% and a ROC of 0.86 on the validation dataset. These results are encouraging and indicate that with further development Deep Learning has the potential to be clinically useful for automated Pneumothorax detection
ZMPCZM016000.11.23 Electrotherapy for pain managementpainezeeman
This document summarizes research on the use of electrotherapy/electrical stimulation for pain management. It discusses two major theories for how electrotherapy relieves pain through gate control and opiate-mediated control. Research studies cited found electrotherapy effective at reducing pain and improving function for chronic musculoskeletal pain, low back pain, and post-operative knee pain. Meta-analyses showed significant decreases in pain from electrical nerve stimulation and reductions in analgesic consumption when using adequate stimulation parameters.
Neurobionics and robotic neurorehabilitationsNeurologyKota
This document discusses neurobionics, robotic neurorehabilitation, and applications of neurobionics. It summarizes key areas including: (1) neurobionics aims to integrate electronics with the nervous system to repair or substitute impaired functions, (2) robotic neurorehabilitation uses robots to assist in rehabilitation processes, and (3) applications of neurobionics include motor interfaces like robotic arms, sensory interfaces like cochlear implants, and treating conditions like epilepsy and Parkinson's disease.
1) A previous study found that repetitive painful stimulation led to decreased pain ratings over 8 days, associated with increased activity in the rACC region.
2) A 1-year follow-up of 10 participants from the original study found pain thresholds had returned to baseline levels. Pain ratings and rACC activity patterns also returned to initial levels.
3) These findings indicate that habituation to pain involves increased activity of the endogenous antinociceptive system, but this is a temporary adaptive process that recedes once painful stimulation ends, though it may last beyond the stimulation period.
This document describes a project aimed at classifying Obstructive Sleep Apnea Syndrome (OSAS) and Major Depressive Disorder (MDD) in patients using machine learning techniques. The project uses a dataset containing neurocognitive measures like EEG signal features and physiologic measures like heart rate and blood pressure. It aims to develop Python routines using QDA classification and RBF neural networks to analyze the data and accurately classify the severity of patients' OSAS, which current medical experts have difficulty doing by manually analyzing multiple data sources. The project is designed in modules and uses software tools like Python, PyQt, and SQLite3 for implementation.
Similar to PainQx -Science Business Development 3.21.16 (20)
1. Alejandro Zamorano
PainQx
VP of Business Development
E: AZamorano@PainQx.com
M: 508.397.8440
O: 617.981.7753
www.painqx.com
Objective Quantification of Pain
2. Overview
PainQx is a med-tech diagnostic/ software company that has developed a next
generation approach to objectively measure pain in humans. The PainQx platform
achieves this by assessing neural activity from a patient’s brain and processing and
decoding the data through proprietary analysis and algorithms. The outputs are scaled
pain and neurological side effects biomarkers that directly correlate to a patient’s pain
state.
The PainQx system is currently being used as a clinical assessment tool by companies
using pain as a primary or secondary clinical end point to measure their drug’s and/or
device’s efficacy.
• Proven results in human clinical setting
• Experienced, world renowned scientific team
• Confirms the actual level of pain in chronic pain patients
• Solution to a vast unmet clinical research need
2
3/21/2016
4. 4
3/21/2016
Problem/Solution in Clinical Trials
PainQx solves the logic chain below by providing:
1. Objective Pain Measurement: Reduces clinical trial costs and risks
2. Scaled Pain and Neurological Side-effect Biomarkers: Improves data, research
outcomes; validates research products
Increased Risk/
Failure Rates of
Pain Drugs and
Treatments
Patient Reporting
Variability &
Subjectivity in
Trial Endpoints
Subjective Pain
Measurement
Large Patient
Sample Size to
smooth out
Variability
Uncertainty in
Reliability of
Clinical Data
Difficulty in
Screening
Placebo
Responders
Difficult to interpret
and determine
pain drug efficacy
Increased Trial
Costs where pain
is measured or an
End point
Issue Creates
Leads to Result
5. Partner Testimonial
“PainQx reduces clinical and regulatory risk. It is based on substantive,
human data and may transform the landscape for pain researchers as
well as patients with pain, by enabling novel and more effective pain
therapeutics to be delivered”. –Yuri Maricich MD MBA, VP of Business
Development and Neurology, Cavion Pharma
5
3/21/2016
6. PainQx assesses well-known, established brain neural network
correlated to pain also known as the “Pain Matrix*”:
Underlying Science
6
3/21/2016
Spinal
Cord
Inputs
Main regions involved in
experiencing pain
Pain Matrix
* Melzack R (Dec 2001). "Pain and the Neuromatrix in the Brain".J Dent Educ 65 (12): 1378–82. PMID 11780656
Brain Regions and Brain Neural
Activity involved with Pain
Captured by EEG & Neuro Imaging
Analysis by PainQx technology
PainQx Proprietary Algorithms decode
the Pain Matrix providing an objective,
replicable method of quantifying pain.
7. 7
3/21/2016
Core Technologies/ Competencies
NYU Brain Research Laboratory Normative Database (BRLND)
PainQx has an exclusive license to the NYU Brain Research Laboratory
Normative Database for use in pain. The BRLND is a 20,000 patient
database of EEG data built over the past 30 years. The BRLND is the
largest of its kind in the world. EEG data has been shown to be culture,
ethnic, and gender free; but not age free.
Electroencephalogram Localization (EEGLx)
EEGLx allows PainQx to process and decode electrical neural activity to
localize activated brain regions
Creates a 3D representation of the brain representing over/under activated
regions relative to norms in the BRLND
Pain Matrix Correlation (PMCx)
PMCx allows PainQx to filter out areas not correlated to the sensation and
perception of pain
Areas of interest are isolated, identified, correlated, and weighted to produce
an objective measurement of a patient’s pain state
8. A normal brain looks different than a brain in
pain as seen using the PainQx technology.
Before
Surgery
After
Surgery
Patient Case Study:
Patient with chronic neuropathic
pain caused by nerve
compression in the spine.
Low Pain High Pain
Data: High & Low Pain State for an
Individual Patient
8
3/21/2016
Top Panel: Patient Before
Surgical Decompression
Over-activated brain regions
relative to norms correlated w/
Pain Matrix. Patient reported pain
score: [4 out of 10]
Bottom Panel: Patient After
Surgical Decompression
Low-activated brain regions
relative to norms correlated w/
Pain Matrix. After surgery, patient
reported pain score: [0 out of 10]
Prichep LS, John ER, Howard B, Merkin H, Hiesiger EM. Evaluation of the pain matrix
using EEG source localization: a feasibility study. Pain Med. 2011;12(8):1241-8.
9. 65 Patient study to compare
PainQx generated score
compared with VAS pain score
Inclusion Criteria:
Male and Female chronic pain
patients ages 19 – 82
All patients have a history of at
least 3 months of pain
Patients were excluded with
neurological or psychiatric
disorders (including drug and
alcohol dependence), head
injuries or skull abnormalities.
Demonstrates ability to clearly
differentiate the VAS scale by as
few as one point gradation
Regression (n=65) QEEG evaluation of Pain States w/ VAS
scores between 0-10: R=0.907, p<0.00001
Data: Generated PainQx Scores vs
Patient Derived VAS Pain Score
9
3/21/2016
VAS Pain Score
Publications: Prichep et al., 2015
1
PainQxScore
10. 10
3/21/2016
PainQx Value Proposition to Partners
Data Generated
PainQx provides quantitative measures of activity in different regions of the brain (Regions of
Interest, ROIs) which are involved in the sensation and perception of pain (i.e. the “Pain Matrix”)
ROIs include: cingulate, insula, dorsolateral prefrontal and thalamic regions
Interpretation of Data
PainQx decodes the activation of the brain structures involved in the sensation and perception
of pain to provide a marker of this pain activation
Changes in these markers can be tracked over time
Normalization of activation of ROIs can provide a quantitative measure of positive response to a drug
These markers can be used to identify optimal dosage levels of a drug
Specific ROIs can be selected to facilitate testing of hypotheses related to drug targets
Extrapolation
Since ROIs include areas related to both sensation and perception, distinctive patterns can be
identified which are more closely related to different aspects of the painful response, which may
include:
Comorbidities (e.g., depression)
Placebo response
11. 11
Process for Generating a Pain Report
Patient EEG is
collected (10 mins
worth of data)
Collection Cloud Server
EEG data is sent to
PQX via secure
cloud
Clean Signal
PQX receives the
file and removes
artifacts
EEGLx
3D image of brain
and its activity is
created
Normative Database (BRLND)
Pain Matrix
Brain regions not
correlated to Pain
Matrix are filtered
out
PainQx Analysis
Activated brain regions are compared to
patients within their age group in the BRLND
to identify abnormal activity compared to the
“Ground State” of the brain
Regions of interest related to the Pain
Matrix are isolated, identified, correlated,
and weighted to produce an objective
measurement of a patient’s pain state
Patient
Pain
Report
3/21/2016
12. 12
3/21/2016
Deliverable
PainQx will issue a report which will contain quantitative measures and images
of the significance of over/under activation in brain Regions of Interest (ROIs):
Provides an initial biomarker for patient status prior to therapeutic intervention
Provides a quantitative method to track the significance of the change in the
biomarker (within specific ROIs) following intervention
Individual Patient and Aggregate Report:
o EEG source localization images superimposed on an MRI atlas showing sequential
slices for the axial view of the brain, and 3 summary images (axial, sagittal and
coronal views of maximum abnormality) (Exhibit 1)
o Table of the mean z-scores for each ROI (where z=1.96, p<0.05), also shown are
standard deviation, minimal and maximal z value (Exhibit 2)
Aggregate Report:
o Group average data can be similarly displayed for sets of patients at baseline and
after intervention (Exhibit 1)
o Significance of the changes for the group can be displayed as t-scores for the
significance of the change and displayed as above, along with tabular values for
the significance of the changes by ROI (Exhibit 3)
13. 13
3/21/2016
Exhibit 1: Customer Deliverable
28 QEEG Images showing activation of pain matrix relative to normative
population. Colors demonstrate neural activity’s standard deviation from
the norm.
Under- Over-
Activation Activation
14. ROIs Brodmann Area Mean z Max z Min z
Posterior Cingulate 23, 29, 30, 31 2.57 3.04 2.01
L Inferior Parietal Lobule 39, 40 2.52 2.94 1.82
Post L Insula 13 2.50 2.77 2.06
L Superior Temporal Gyrus 22 2.49 2.90 1.95
L Parahippocampal Gyrus Hippocampus, Amygdala 2.45 2.70 2.27
L Precuneus 7 2.35 2.90 1.81
Post R Insula 13 2.30 2.80 1.56
Thalamus Pulvinar 2.29 2.56 2.05
R Inferior Parietal Lobule 39, 40 2.29 2.95 1.40
R Parahippocampal Gyrus Hippocampus, Amygdala 2.28 2.50 2.05
Table of quantitative values of mean z-score for each of the listed ROIs from
the previous image. A z-value ≥1.96 (red font) is significantly different from
normal at the p<0.05 level.
Exhibit 2: Customer Deliverable
Most
significant
ROIs
Brodmann
area of the
ROIs
Mean
Z-Score
• ROIs can include all regions, only significant regions, or specific regions of focus and can be
configured for particular use
• Intervention can be evaluated quantitatively as magnitude of changes in Z-scores – where
decreases indicate movement toward normal or positive change from baseline in the patient
Maximum
and
Minimum
Z-Score
14
3/21/2016
15. 15
3/21/2016
Exhibit 3: Customer Deliverable
Group analyses can also be made, statistically comparing Baseline (Pre-Drug)
and Post-Drug measures in each ROI
8.6 Hz
Region BA t-value p-vlaue
R Precuneus 7 -147.068 <0.0001
L Precuneus 7 -136.318 <0.0001
Mid Cingulate 31 -128.441 <0.0001
L DLPFC 46 -49.3934 <0.0001
Posterior Cingulate 30 -45.2018 <0.0001
Anterior Cingulate 25 -42.4683 <0.0001
R DLPFC 46 -4.74674 <0.0001
L Insula 13 64.6416 <0.0001
R Insula 13 38.9874 <0.0001
L Parahippocampus Hippocampus, Amygdala 29.4841 <0.0001
R Parahippocampus Hippocampus, Amygdala 27.0849 <0.0001
R Orbital-frontal cortex 11, 47 23.3455 <0.0001
Thalamus Anterior nucleus 17.1622 <0.0001
• Significant negative
t-values show
decreases in regions
over-activation at
baseline
• Significant positive
t-values show
increases in
activation in regions
under-activated at
baseline