Precise Image-level Localization of Intracranial Hemorrhage on Head CT Scans with Deep Learning Models Trained on Study-level Labels.

"Just Accepted" papers have undergone full peer review and have been accepted for publication in Radiology: Artificial Intelligence. This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content. Purpose To develop a highly generalizable weakly supervised model to automatically detect and localize image- level intracranial hemorrhage (ICH) using study-level labels. Materials and Methods In this retrospective study, the proposed model was pretrained on the image-level RSNA dataset and fine-tuned on a local dataset using attention-based bidirectional long-short-term memory networks. This local training dataset included 10,699 noncontrast head CT scans from 7469 patients with ICH study-level labels extracted from radiology reports. Model performance was compared with that of two senior neuroradiologists on 100 random test scans using the McNemar test, and its generalizability was evaluated on an external independent dataset. Results The model achieved a positive predictive value (PPV) of 85.7% (95% CI: [84.0%, 87.4%]) and an AUC of 0.96 (95% CI: [0.96, 0.97]) on the held-out local test set (n = 7243, 3721 female) and 89.3% (95% CI: [87.8%, 90.7%]) and 0.96 (95% CI: [0.96, 0.97]), respectively, on the external test set (n = 491, 178 female). For 100 randomly selected samples, the model achieved performance on par with two neuroradiologists, but with a significantly faster (P < .05) diagnostic time of 5.04 seconds per scan (versus 86 seconds and 22.2 seconds for the two neuroradiologists, respectively). The model's attention weights and heatmaps visually aligned with neuroradiologists' interpretations. Conclusion The proposed model demonstrated high generalizability and high PPVs, offering a valuable tool for expedited ICH detection and prioritization while reducing false-positive interruptions in radiologists' workflows. ©RSNA, 2024.

Temporal Changes on Postgadolinium MR Vessel Wall Imaging Captures Enhancement Kinetics of Intracranial Atherosclerotic Plaques and Aneurysms.

BACKGROUND AND PURPOSE: Analysis of vessel wall contrast kinetics (ie, wash-in/washout) is a promising method for the diagnosis and risk-stratification of intracranial atherosclerotic disease plaque (ICAD-P) and the intracranial aneurysm walls (IA-W). We used black-blood MR imaging or MR vessel wall imaging to evaluate the temporal relationship of gadolinium contrast uptake kinetics in ICAD-Ps and IA-Ws compared with normal anatomic reference structures. MATERIALS AND METHODS: Patients with ICAD-Ps or IAs who underwent MR vessel wall imaging with precontrast, early postcontrast (5-15 minutes), and delayed postcontrast (20-30 minutes) 3D T1-weighted TSE sequences were retrospectively studied. ROIs of a standardized diameter (2 mm) were used to measure the signal intensities of the cavernous sinus, pituitary infundibulum, temporalis muscle, and choroid plexus. Point ROIs were used for ICAD-Ps and IA-Ws. All ROI signal intensities were normalized to white matter signal intensity obtained using ROIs of 10-mm diameter. Measurements were acquired on precontrast, early postcontrast, and delayed postcontrast 3D T1 TSE sequences for each patient. RESULTS: Ten patients with 17 symptomatic ICAD-Ps and 30 patients with 34 IA-Ws were included and demonstrated persisting contrast uptake (P < .001) of 7.21% and 10.54% beyond the early phase (5-15 minutes postcontrast) and in the delayed phase (20-30 minutes postcontrast) on postcontrast MR vessel wall imaging. However, normal anatomic reference structures including the pituitary infundibulum and cavernous sinus demonstrated a paradoxical contrast washout in the delayed phase. In both ICAD-Ps and IA-Ws, the greatest percentage of quantitative enhancement (>70%-90%) occurred in the early phase of postcontrast imaging, consistent with the rapid contrast uptake kinetics of neurovascular pathology. CONCLUSIONS: Using standard MR vessel wall imaging techniques, our results demonstrate the effects of gadolinium contrast uptake kinetics in ICAD-Ps and IA-Ws with extended accumulating enhancement into the delayed phase (> 15 minutes) as opposed to normal anatomic reference structures that conversely exhibit decreasing enhancement. Because these relative differences are used to assess qualitative patterns of ICAD-P and IA-W enhancement, our findings highlight the importance of standardizing acquisition time points and MR vessel wall imaging protocols to interpret pathologic enhancement for the risk stratification of cerebrovascular pathologies.

Bivalirudin as a substitute for heparin in neurointervention for patients with heparin-induced thrombocytopenia.

OBJECTIVES: Heparin-induced thrombocytopenia is a known complication of heparin exposure with potentially life-threatening sequelae. Direct thrombin inhibitors can be substituted for heparin in patients with heparin-induced thrombocytopenia that require anticoagulation. However, the use of direct thrombin inhibitors as a substitute for heparin has not been widely reported in the neuroendovascular literature. MATERIALS AND METHODS: Here we report the first use of the direct thrombin inhibitor bivalirudin in a neuroendovascular procedure as a substitute for heparin in a patient with a ruptured pseudoaneurysm and heparin-induced thrombocytopenia, and review the literature on the use of bivalirudin and argatroban for such patients. RESULTS: Bivalirudin was safely and effectively used in the case reported, with no thrombotic or hemorrhagic complications. Our literature review revealed a paucity of studies on the use of heparin alternatives, including bivalirudin, in neuroendovascular procedures in patients with heparin-induced thrombocytopenia. CONCLUSIONS: Heparin-induced thrombocytopenia is an important iatrogenic disease process in patients undergoing neuroendovascular procedures, and developing protocols to diagnose and manage heparin-induced thrombocytopenia is important for healthcare systems. While further research needs to be done to establish the full range of anticoagulation options to substitute for heparin, our case indicates bivalirudin as a potential candidate.

MR Vessel Wall Imaging for Atherosclerosis and Vasculitis.

Conventional imaging modalities, such as computed tomography angiography, MR angiography, transcranial Doppler ultrasonography, and digital subtraction angiography, are utilized in evaluating intraluminal or intravascular pathology of the intracranial vessels. Limitations of luminal imaging techniques can lead to inaccurate diagnosis, evaluation, and risk stratification, as many cerebrovascular pathologies contain an extrinsic vessel wall component. Furthermore, vessel wall imaging can provide information regarding extent, treatment response, and biopsy targets for vasculitis cases. Overall, while vessel wall imaging can provide robust data regarding intracranial pathologies, further prospective, multicenter studies are required to improve diagnostic application and accuracy.

Background Subtraction Angiography with Deep Learning Using Multi-frame Spatiotemporal Angiographic Input.

Catheter Digital Subtraction Angiography (DSA) is markedly degraded by all voluntary, respiratory, or cardiac motion artifact that occurs during the exam acquisition. Prior efforts directed toward improving DSA images with machine learning have focused on extracting vessels from individual, isolated 2D angiographic frames. In this work, we introduce improved 2D + t deep learning models that leverage the rich temporal information in angiographic timeseries. A total of 516 cerebral angiograms were collected with 8784 individual series. We utilized feature-based computer vision algorithms to separate the database into "motionless" and "motion-degraded" subsets. Motion measured from the "motion degraded" category was then used to create a realistic, but synthetic, motion-augmented dataset suitable for training 2D U-Net, 3D U-Net, SegResNet, and UNETR models. Quantitative results on a hold-out test set demonstrate that the 3D U-Net outperforms competing 2D U-Net architectures, with substantially reduced motion artifacts when compared to DSA. In comparison to single-frame 2D U-Net, the 3D U-Net utilizing 16 input frames achieves a reduced RMSE (35.77 ± 15.02 vs 23.14 ± 9.56, p < 0.0001; mean ± std dev) and an improved Multi-Scale SSIM (0.86 ± 0.08 vs 0.93 ± 0.05, p < 0.0001). The 3D U-Net also performs favorably in comparison to alternative convolutional and transformer-based architectures (U-Net RMSE 23.20 ± 7.55 vs SegResNet 23.99 ± 7.81, p < 0.0001, and UNETR 25.42 ± 7.79, p < 0.0001, mean ± std dev). These results demonstrate that multi-frame temporal information can boost performance of motion-resistant Background Subtraction Deep Learning algorithms, and we have presented a neuroangiography domain-specific synthetic affine motion augmentation pipeline that can be utilized to generate suitable datasets for supervised training of 3D (2d + t) architectures.

Unplanned readmission after carotid stenting versus endarterectomy: analysis of the United States Nationwide Readmissions Database.

BACKGROUND: Hospital readmissions are costly and reflect negatively on care delivered. OBJECTIVE: To have a better understanding of unplanned readmissions after carotid revascularization, which might help to prevent them. METHODS: The Nationwide Readmissions Database was used to determine rates and reasons for unplanned readmission following carotid endarterectomy (CEA) and carotid artery stenting (CAS). Trends were assessed by annual percent change, modified Poisson regression was used to estimate risk ratios (RR) for readmission, and propensity scores were used to match cohorts. RESULTS: Analysis yielded 522 040 asymptomatic and 55 485 symptomatic admissions for carotid revascularization between 2010 and 2015. Higher 30-day readmission rates were noted after CAS versus CEA in both symptomatic (9.1% vs 7.7%, p<0.001) and asymptomatic (6.8% vs 5.7%, p<0.001) patients. Readmission rates trended lower over time, significantly so for 90-day readmissions in symptomatic patients undergoing CEA. The most common cause for 30-day readmission was stroke in both symptomatic (5.5%) and asymptomatic (3.9%) patients. Factors associated with a higher risk of readmission included age over 80; male gender; Medicaid health insurance; and increases in severity of illness, mortality risk, and comorbidity indices. Analysis of matched cohorts showed that CAS had higher readmission than CEA (RR=1.14 (95% CI 1.06 to 1.22); p<0.001) only in asymptomatic patients. Adverse events during initial admission which predicted 30-day readmission included acute renal failure and acute respiratory failure in asymptomatic patients; hematoma and cardiac events were additional predictive adverse events in symptomatic patients. CONCLUSIONS: Readmission is not uncommon after carotid revascularization, occurs more often after CAS, and is predicted by baseline factors and by preventable adverse events at initial admission.

Holohemispheric Prostate Carcinoma Dural Metastasis Mimicking Subdural Hematoma: Case Report and Review of the Literature.

Prostate carcinomas are the most common malignancy to metastasize to the dura. These metastases can commonly mimic subdural hematomas and may similarly present with brain compression. The optimal management and outcomes after surgical management are not well characterized. We present a case of prostate carcinoma metastatic to the dura that was initially thought to be a large isodense subdural hematoma and was treated with surgical decompression. We also review the literature regarding prostate dural metastases mimicking subdural hematomas and discuss the relevant imaging findings, treatments, and outcomes. Dural metastasis should be considered when a patient with known metastatic prostate cancer presents with imaging evidence of a subdural mass.

Case Report: Anterior Spinal Cord Ischemia Following Embolization of Cerebellar Arteriovenous Malformation: An Illustrative Case and Review of Spinal Cord Vascular Anatomy.

Spinal cord ischemia (SCI) is a rare entity with high mortality and morbidity which can arise from causes such as atherosclerosis, aortic dissection or aneurysm, thromboembolic events or systemic hypotension, and is a potential complication of spinal surgery. Published literature contains very few reports of SCI as a complication of intracranial interventions, highlighting the uncommon nature of SCI in these circumstances. We report the occurrence of anterior SCI in a 69-year-old patient following successful embolization of a cerebellar arteriovenous malformation (AVM), marked by upper extremity weakness, lower extremity paraplegia, loss of bladder and bowel control, and hypercapnic respiratory failure requiring mechanical ventilation. Magnetic resonance imaging (MRI) demonstrated upper cervical diffusion restriction and T2/STIR hyperintensity. Unusually, SCI occurred in this case without intraprocedural catheter wedging or obvious flow limitation, prolonged procedure time, hypercoagulable state, or general hypotension. We review previous cases in the literature as well as spinal cord vascular anatomy, and discuss the possible etiologies of this complication. Spinal cord ischemia could be a very rare complication of neuroendovascular procedures even in the absence of warning signs and should be carefully evaluated in patients with suspected neurologic symptoms after such procedures.

Pipeline embolization of cerebral aneurysms in pediatric patients: combined systematic review of patient-level data and multicenter retrospective review.

OBJECTIVE: Cerebral aneurysms in the pediatric population are rare and optimal treatment strategies are not as well characterized as in adults. The Pipeline embolization device (PED) is an endoluminal flow diverter that is commonly used to treat aneurysms in adults, but experience with this device in children is limited. The authors sought to further characterize PED use and outcomes in this specific population by performing both a systematic review of patient-level data from studies reporting the use of the PED to treat pediatric aneurysms and a retrospective review of their experience. METHODS: A systematic review of the PubMed, Embase, and Scopus databases was performed to identify studies reporting the use of the PED in pediatric patients (age ≤ 18 years). Disaggregated data regarding demographics, aneurysm characteristics, treatment, and outcomes were collected. Retrospective data from the authors' two institutions were also included. RESULTS: Thirty studies comprising patient-level data on 43 pediatric patients with 47 aneurysms were identified. An additional 9 patients with 9 aneurysms were included from the authors' institutions for a total of 52 patients with 56 aneurysms. The mean patient age was 11.1 years. Presentations included aneurysm rupture (17.3%) and symptomatic mass effect (23.1%). Aneurysms were located in the anterior circulation in 55.4% of cases, and 73.2% were described as nonsaccular. Imaging follow-up was available for 89.3% with a mean follow-up of 13.3 months. Aneurysm occlusion was reported in 75%, with 1 case each (1.8%) demonstrating significant in-stent stenosis and parent vessel occlusion. Clinical follow-up was reported in 90.4% with a mean follow-up of 14.7 months. Good functional outcomes (modified Rankin Scale score of 0-1 or Glasgow Outcome Scale score of 5) were reported in 65.4% of the total population. Two major complications were reported, including 1 death. CONCLUSIONS: Despite substantial differences in aneurysm location and type between published pediatric and adult patient populations treated with the PED, the use of the PED in the pediatric population appears to be safe. While the short-term effectiveness is also similar to that of adults, additional studies are needed to further characterize the long-term outcomes and better define the use of this device in pediatric patients.

Direct percutaneous transcarotid approach for embolization of dural sinus malformation in a premature neonate: A case report.

CONCLUSIONS: Direct percutaneous TCAA can be a feasible access for intracranial interventions in neonates with LBW and unavailable UA access.

DeepCOVID-XR: An Artificial Intelligence Algorithm to Detect COVID-19 on Chest Radiographs Trained and Tested on a Large U.S. Clinical Data Set.

Background There are characteristic findings of coronavirus disease 2019 (COVID-19) on chest images. An artificial intelligence (AI) algorithm to detect COVID-19 on chest radiographs might be useful for triage or infection control within a hospital setting, but prior reports have been limited by small data sets, poor data quality, or both. Purpose To present DeepCOVID-XR, a deep learning AI algorithm to detect COVID-19 on chest radiographs, that was trained and tested on a large clinical data set. Materials and Methods DeepCOVID-XR is an ensemble of convolutional neural networks developed to detect COVID-19 on frontal chest radiographs, with reverse-transcription polymerase chain reaction test results as the reference standard. The algorithm was trained and validated on 14 788 images (4253 positive for COVID-19) from sites across the Northwestern Memorial Health Care System from February 2020 to April 2020 and was then tested on 2214 images (1192 positive for COVID-19) from a single hold-out institution. Performance of the algorithm was compared with interpretations from five experienced thoracic radiologists on 300 random test images using the McNemar test for sensitivity and specificity and the DeLong test for the area under the receiver operating characteristic curve (AUC). Results A total of 5853 patients (mean age, 58 years ± 19 [standard deviation]; 3101 women) were evaluated across data sets. For the entire test set, accuracy of DeepCOVID-XR was 83%, with an AUC of 0.90. For 300 random test images (134 positive for COVID-19), accuracy of DeepCOVID-XR was 82%, compared with that of individual radiologists (range, 76%-81%) and the consensus of all five radiologists (81%). DeepCOVID-XR had a significantly higher sensitivity (71%) than one radiologist (60%, P < .001) and significantly higher specificity (92%) than two radiologists (75%, P < .001; 84%, P = .009). AUC of DeepCOVID-XR was 0.88 compared with the consensus AUC of 0.85 (P = .13 for comparison). With consensus interpretation as the reference standard, the AUC of DeepCOVID-XR was 0.95 (95% CI: 0.92, 0.98). Conclusion DeepCOVID-XR, an artificial intelligence algorithm, detected coronavirus disease 2019 on chest radiographs with a performance similar to that of experienced thoracic radiologists in consensus. © RSNA, 2020 Supplemental material is available for this article. See also the editorial by van Ginneken in this issue.

Magnetization Transfer Ratio and Morphometrics of the Spinal Cord Associates with Surgical Recovery in Patients with Degenerative Cervical Myelopathy.

OBJECTIVES: We assessed the prognostic value of the preoperative magnetization transfer ratio (MTR) and morphometrics of the spinal cord in patients with degenerative cervical myelopathy (DCM) in a longitudinal cohort study. METHODS: Thirteen subjects with DCM underwent 3T magnetization transfer imaging. The MTR was calculated for the spinal cord regions and specific white matter tracts. Morphometric measures were extracted. Clinical (modified Japanese Orthopaedics Association [mJOA] and Nurick scale scores) and health-related quality of life scores were assessed before and after cervical decompression surgery. The association between the magnetic resonance imaging (MRI) metrics and postoperative recovery was assessed (Spearman's correlation). Receiver operating characteristics were used to assess the accuracy of MRI metrics in identifying ≥50% recovery in function. RESULTS: Preoperative anterior cord MTRs were associated with recovery in mJOA scores (ρ = 0.608; P = 0.036; area under the curve [AUC], 0.66). Preoperative lateral cord MTR correlated with the neck disability index (ρ = 0.699; P = 0.011) and pain interference scale (ρ = 0.732; P = 0.007). Preoperative rubrospinal tract MTR was associated with mJOA score recovery (ρ = 0.573; P = 0.041; AUC, 0.86). Preoperative corticospinal tract and reticulospinal MTRs were related to recovery in pain interference scores (ρ = 0.591; P = 0.033; and ρ = 0.583; P = 0.035, respectively). Eccentricity of the cord was associated with Nurick scores (ρ = 0.606; P = 0.028) and mJOA scores (ρ = 0.651; P = 0.025; AUC, 0.92). CONCLUSIONS: Preoperative MTR and eccentricity measurements of the spinal cord have prognostic value in assessing the response to surgery and recovery in patients with DCM. Advanced MRI and atlas-based postprocessing techniques can inform interventions and advance the healthcare received by patients with DCM.

Conservative Management and Natural History of Ruptured Basilar Perforator Artery Aneurysms: Two Cases and Literature Review.

BACKGROUND: Basilar perforator aneurysms are rare causes of subarachnoid hemorrhage and their natural history is poorly characterized. Although various treatment strategies have been reported, conservative management is an option that has been associated with a high likelihood of spontaneous resolution. CASE DESCRIPTION: Here we present 2 cases of subarachnoid hemorrhage, 1 diffuse and the other perimesencephalic, due to small ruptured basilar perforator artery aneurysms. These aneurysms were only identified after repeat angiography. Conservative management with serial imaging was pursued. Both patients did well clinically and repeat imaging demonstrated spontaneous resolution of the ruptured aneurysms. We also provide a literature review of ruptured basilar perforator aneurysms, showing a ~10% re-rupture rate within the early post-rupture period but otherwise a high rate of spontaneous resolution. CONCLUSIONS: Although basilar perforator aneurysms can re-rupture, there is also a high likelihood of spontaneous resolution. Given the challenges of treatment, conservative management is an option that can be considered.

Texture analysis of placental MRI: can it aid in the prenatal diagnosis of placenta accreta spectrum?

PURPOSE: To determine if texture analysis can differentiate placenta accreta spectrum (PAS) from normal placenta on MRI. METHODS: We performed retrospective image analysis of 80 patients, comprised of 46 patients with PAS and 34 patients without PAS. Histopathology was used as the reference standard. Sagittal single shot fast spin echo T2-weighted MRI sequences acquired from a single institution were analyzed. Placental heterogeneity was quantified using in-house software on a Matlab platform, including the standard deviation of pixel intensity, coefficient of variation, gray-level co-occurrence matrices (GLCM), histogram-oriented gradients (HOG), and fractal analysis with box sizes from 2 to 512. Two-tailed unpaired Student's t test was used with statistical significance of p < 0.05. RESULTS: PAS was associated with higher values for standard deviation of pixel intensity and fractal analysis at every box size. Fractal analysis at box sizes 256 (p = 0.011) and 32 (p = 0.021), and standard deviation of pixel intensity (p = 0.023) were the most statistically significant. Fractal values at box size 256 for PAS was 0.13 versus 0.090 for patients without PAS, while standard deviation of pixel intensity was 3.7 for PAS versus 2.5 for patients without PAS. No statistically significant association between PAS and GLCM, coefficient of variation, and HOG was found. CONCLUSION: Statistically significant differences were found between normal and abnormal groups using standard deviation of pixel intensity and fractal analysis.

Machine Learning for the Prediction of Cervical Spondylotic Myelopathy: A Post Hoc Pilot Study of 28 Participants.

BACKGROUND: Cervical spondylotic myelopathy (CSM) severity and presence of symptoms are often difficult to predict based simply on clinical imaging alone. Similarly, improved machine learning techniques provide new tools with immense clinical potential. METHODS: A total of 14 patients with CSM and 14 controls underwent imaging of the cervical spine. Two different artificial neural network models were trained; 1) to predict CSM diagnosis; and 2) to predict CSM severity. Model 1 consisted of 6 inputs including 3 common imaging scales for the evaluation of cord compression, alongside 3 objective magnetic resonance imaging measurements. The outcome for model 1 was binary to predict CSM diagnosis. Model 2 consisted of 23 input variables derived from probabilistic volume mapping measurements of white matter tracts in the region of compression. The outcome of model 2 was linear, to predict the modified Japanese Orthopedic Association (mJOA) score. RESULTS: Model 1 was used in predicting CSM. The mean cross-validated accuracy of the trained model was 86.50% (95% confidence interval, 85.16%-87.83%) with a median accuracy of 90.00%. Area under the curve (AUC) was calculated for each repetition. Average AUC for each repetition was 0.947 with a median AUC of 1.0. Average sensitivity, specificity, positive predictive value, and negative predictive value were 90.25%, 85.05%, 81.58%, and 91.94%, respectively. Model 2 was used in modeling mJOA. The mJOA model predicted scores, with a mean and median error of -0.29 mJOA points and -0.08 mJOA points, respectively, mean error per batch was 0.714 mJOA points. CONCLUSIONS: Machine learning provides a promising method for prediction, diagnosis, and even prognosis in patients with CSM.

Non-centered spike-triggered covariance analysis reveals neurotrophin-3 as a developmental regulator of receptive field properties of ON-OFF retinal ganglion cells.

The functional separation of ON and OFF pathways, one of the fundamental features of the visual system, starts in the retina. During postnatal development, some retinal ganglion cells (RGCs) whose dendrites arborize in both ON and OFF sublaminae of the inner plexiform layer transform into RGCs with dendrites that monostratify in either the ON or OFF sublamina, acquiring final dendritic morphology in a subtype-dependent manner. Little is known about how the receptive field (RF) properties of ON, OFF, and ON-OFF RGCs mature during this time because of the lack of a reliable and efficient method to classify RGCs into these subtypes. To address this deficiency, we developed an innovative variant of Spike Triggered Covariance (STC) analysis, which we term Spike Triggered Covariance - Non-Centered (STC-NC) analysis. Using a multi-electrode array (MEA), we recorded the responses of a large population of mouse RGCs to a Gaussian white noise stimulus. As expected, the Spike-Triggered Average (STA) fails to identify responses driven by symmetric static nonlinearities such as those that underlie ON-OFF center RGC behavior. The STC-NC technique, in contrast, provides an efficient means to identify ON-OFF responses and quantify their RF center sizes accurately. Using this new tool, we find that RGCs gradually develop sensitivity to focal stimulation after eye opening, that the percentage of ON-OFF center cells decreases with age, and that RF centers of ON and ON-OFF cells become smaller. Importantly, we demonstrate for the first time that neurotrophin-3 (NT-3) regulates the development of physiological properties of ON-OFF center RGCs. Overexpression of NT-3 leads to the precocious maturation of RGC responsiveness and accelerates the developmental decrease of RF center size in ON-OFF cells. In summary, our study introduces STC-NC analysis which successfully identifies subtype RGCs and demonstrates how RF development relates to a neurotrophic driver in the retina.

Extracellular stimulation of mouse retinal ganglion cells with non-rectangular voltage-controlled waveforms.

Neural prostheses rely upon electric stimulation to control neural activity. However, electrode corrosion and tissue damage may result from the injection of high charge densities. During electrical stimulation with traditional voltage-controlled square-wave pulses, the current density distribution on the surface of the stimulating electrode is highly nonuniform, with the highest current densities located at the edge of disk-shaped electrodes. Current density is implicated in tissue damage and electrode corrosion because it determines the charge density distribution. Through recent computer modeling work, we have found that Gaussian and sinusoidal stimulus waveforms produce a current density distribution that is significantly more uniform than the one produced by square-wave pulses. In this manner, these non-rectangular waveforms reduce the peak current densities without decreasing the efficacy of the neural stimulus. In the present work, we utilize an in vitro mouse retinal preparation to compare the same set of alternative stimulus waveforms. The -1V amplitude voltage-controlled stimuli were delivered through 20 microm diameter titanium nitride electrodes. Importantly, when normalized for the amount of injected charge, the data demonstrate that each waveform is similarly effective at eliciting a neural response. Also, the suprathreshold Gaussian and sinusoidal waveforms possessed much lower peaks in current. For this reason, these non-rectangular waveforms may be useful in reducing electrode corrosion and tissue damage.

Allelic variance between GRM6 mutants, Grm6nob3 and Grm6nob4 results in differences in retinal ganglion cell visual responses.

An electroretinogram (ERG) screen identified a mouse with a normal a-wave but lacking a b-wave, and as such it was designated no b-wave3 (nob3). The nob3 phenotype mapped to chromosome 11 in a region containing the metabotropic glutamate receptor 6 gene (Grm6). Sequence analyses of cDNA identified a splicing error in Grm6, introducing an insertion and an early stop codon into the mRNA of affected mice (designated Grm6(nob3)). Immunohistochemistry of the Grm6(nob3) retina showed that GRM6 was absent. The ERG and visual behaviour abnormalities of Grm6(nob3) mice are similar to Grm6(nob4) animals, and similar deficits were seen in compound heterozygotes (Grm6(nob4/nob3)), indicating that Grm6(nob3) is allelic to Grm6(nob4). Visual responses of Grm6(nob3) retinal ganglion cells (RGCs) to light onset were abnormal. Grm6(nob3) ON RGCs were rarely recorded, but when they were, had ill-defined receptive field (RF) centres and delayed onset latencies. When Grm6(nob3) OFF-centre RGC responses were evoked by full-field stimulation, significantly fewer converted that response to OFF/ON compared to Grm6(nob4) RGCs. Grm6(nob4/nob3) RGC responses verified the conclusion that the two mutants are allelic. We propose that Grm6(nob3) is a new model of human autosomal recessive congenital stationary night blindness. However, an allelic difference between Grm6(nob3) and Grm6(nob4) creates a disparity in inner retinal processing. Because the localization of GRM6 is limited to bipolar cells in the On pathway, the observed difference between RGCs in these mutants is likely to arise from differences in their inputs.

Incorporation of the electrode-electrolyte interface into finite-element models of metal microelectrodes.

An accurate description of the electrode-electrolyte interfacial impedance is critical to the development of computational models of neural recording and stimulation that aim to improve understanding of neuro-electric interfaces and to expedite electrode design. This work examines the effect that the electrode-electrolyte interfacial impedance has upon the solutions generated from time-harmonic finite-element models of cone- and disk-shaped platinum microelectrodes submerged in physiological saline. A thin-layer approximation is utilized to incorporate a platinum-saline interfacial impedance into the finite-element models. This approximation is easy to implement and is not computationally costly. Using an iterative nonlinear solver, solutions were obtained for systems in which the electrode was driven at ac potentials with amplitudes from 10 mV to 500 mV and frequencies from 100 Hz to 100 kHz. The results of these simulations indicate that, under certain conditions, incorporation of the interface may strongly affect the solutions obtained. This effect, however, is dependent upon the amplitude of the driving potential and, to a lesser extent, its frequency. The solutions are most strongly affected at low amplitudes where the impedance of the interface is large. Here, the current density distribution that is calculated from models incorporating the interface is much more uniform than the current density distribution generated by models that neglect the interface. At higher potential amplitudes, however, the impedance of the interface decreases, and its effect on the solutions obtained is attenuated.

A time domain finite element model of extracellular neural stimulation predicts that non-rectangular stimulus waveforms may offer safety benefits.

Efforts to understand and model the process of extracellular neural electric stimulation have been driven by the desire to intelligently design neural prostheses in order to minimize tissue damage and to maximize the success for stimulating targeted neural structures. Tissue damage and electrode corrosion have been associated with high charge density, which is the integral of the current density passing through the electrode surface over the duration of the stimulus. Importantly, the current density distribution on the surface of stimulating electrodes can be extremely nonuniform, especially when high voltages or frequencies cause a decrease in the electrode-electrolyte impedance. Large current densities are found locally in regions of high curvature, such as the edge of a disk electrode or the tip of a conical electrode. We use a time domain finite element model of a platinum disk electrode and a simplified retinal ganglion cell to explore the potential for Gaussian and sinusoidal voltage-controlled stimulus waveforms to reduce the nonuniformity of the current densities on the electrode surface while maintaining stimulation efficacy. We model an overpotential-dependent electrode-electrolyte interfacial impedance consistent with the platinum-saline interface. An excitable cell membrane is incorporated using the Fohlmeister-Coleman-Miller model of a retinal ganglion cell. Both the electrode-electrolyte interface and the cell membrane were incorporated into the finite element model using a thin layer approximation. All simulations were performed in the COMSOL Multiphysics modeling environment. Rectangular stimulus waveforms were compared to waveforms of Gaussian and sinusoidal shapes. The results suggest that Gaussian and Sinusoidal waveforms may significantly decrease the nonuniformity of the current density distribution while retaining stimulation efficacy.