Explaining neural scaling laws.

The population loss of trained deep neural networks often follows precise power-law scaling relations with either the size of the training dataset or the number of parameters in the network. We propose a theory that explains the origins of and connects these scaling laws. We identify variance-limited and resolution-limited scaling behavior for both dataset and model size, for a total of four scaling regimes. The variance-limited scaling follows simply from the existence of a well-behaved infinite data or infinite width limit, while the resolution-limited regime can be explained by positing that models are effectively resolving a smooth data manifold. In the large width limit, this can be equivalently obtained from the spectrum of certain kernels, and we present evidence that large width and large dataset resolution-limited scaling exponents are related by a duality. We exhibit all four scaling regimes in the controlled setting of large random feature and pretrained models and test the predictions empirically on a range of standard architectures and datasets. We also observe several empirical relationships between datasets and scaling exponents under modifications of task and architecture aspect ratio. Our work provides a taxonomy for classifying different scaling regimes, underscores that there can be different mechanisms driving improvements in loss, and lends insight into the microscopic origin and relationships between scaling exponents.

Assessment of Communication Skills in MBBS Interns with Objective Structured Video Examination (OSVE).

Background: Medical graduates enter work force with substantial knowledge but are they prepared for practice in diverse settings? To train medical interns in communication skills using Demonstration Observation Assistance and Performance (DOAP) and to assess communication skills in medical interns using objective structured video examination (OSVE). Methods and Materials: DOAP sessions of all 27 interns posted in Community Medicine and Pediatrics departments during the duration of the study were conducted in six clinical scenarios followed by two OSVEs. Methods and Material: Interns performed selected activity independently twice for assessments. Videos submitted by interns were assessed by Gap Kalamazoo assessment forms (OSVE). Data analysis was done by appropriate methods. Percentages and proportions and calculation of median scores with interquartile ranges. Results: Seventy one point four percent of faculty and 96.3% of interns were satisfied with module implementation. More than = 57.8% marks were scored by 94.45% interns in self-assessment and faculty assessment and by 100% interns in patient assessment. Median scores of all second assessments were higher. Conclusions: DOAP method is good for training of interns in communication skills. OSVE may be an appropriate assessment tool as it leads to progression in learning.

Compliance With the Royal College of Radiologists Guideline for Actionable Reporting and Its Impact on Patient Care: A Retrospective Analysis of Reporting Practices From a Major Trauma Center.

Introduction Prompt diagnosis forms the mainstay of management of any patient arriving at the hospital. In developed settings, apart from clinical assessment, imaging in the form of computed tomography (CT) scan plays a vital role in arriving at the patient diagnosis. The reporting should follow pre-defined Royal College of Radiologists (RCR) standards to improve the quality of the diagnostic process. Objectives To identify the compliance of reporting as per the RCR standards for the communication of radiological reports and fail-safe alert notification. Materials and methods A retrospective review of body CT scans was done in two cycles within a span of three months. A total of 100 randomized scans were assessed in each cycle, both from the A&E (accident and emergency) and inpatients. Normal scans and outpatient scans were excluded from the study. Data were collected using the online portal (CRIS) and statistical analysis was performed. Results After the first cycle of the audit, 95 reports out of 100 met the standard RCR criteria. After the second cycle, 97 reports met the criteria of the audit. One inpatient scan and two A&E reports did not meet the specified criteria in the second cycle. Conclusion After the two cycles of the audit carried out over three months, we were able to achieve almost 97% of reporting standards as compared to 95% obtained previously through a quality improvement project and create awareness.

Prevalence of High-Risk Pregnancy Among Pregnant Women Attending Antenatal Care Camps in Primary Health Centres in Kinaye and Vantamuri and Their Sub-Centres.

Introduction Identifying pregnancy-related complications and high-risk pregnancies early on and effectively managing care for these pregnant women through a holistic approach guided by the main objectives of antenatal care (ANC) and efficient, good-quality health care through ANC services can reduce the risk of pregnancy-related complications, being favourable for both mother and child. This study was intended to identify the percentage of pregnant women in high-risk groups attending ANC clinics. Objective This study's aim was to understand the prevalence of high-risk pregnancies in women attending ANC camps in Kinaye and Vantamuri villages in Belagavi, India. Methods A community-based cross-sectional study was carried out in a primary healthcare (PHC) setting that included all pregnant women attending antenatal camps and residing in the same areas. With a total study sample size of 200, consisting of pregnant women attending ANC camps in Kinaye and Vantamuri PHC, the data were collected using a pre-tested, pre-designed questionnaire after obtaining written informed consent from the participating pregnant women. Results The data were analysed using a chi-square test to uncover the relationship between socio-demographics, obstetric history, medical variables, and high-risk pregnancy. The results demonstrated that 48.5% of women attending ANC clinics were in the high-risk pregnancy group. Conclusion Pregnancy outcomes are greatly affected by a woman's socio-demographic, obstetric, and medical variables. The WHO recommends a minimum of four ANC visits to safely identify and mitigate the risks and complications of high-risk pregnancy to ensure positive outcomes for both mothers and children.

A Transformer-Based Model for Evaluation of Information Relevance in Online Social-Media: A Case Study of Covid-19 Media Posts.

Online social media has become a major source of information gathering for a huge section of society. As the amount of information flows in online social media is enormous but on the other hand, the fact-checking sources are limited. This shortfall of fact-checking gives birth to the problem of misinformation and disinformation in the case of the truthfulness of facts on online social media which can have serious effects on the wellbeing of society. This problem of misconception becomes more rapid and critical when some events like the recent outbreak of Covid-19 happen when there is no or very little information is available anywhere. In this scenario, the identification of the content available online which is mostly propagated from person to person and not by any governing authority is very needed at the hour. To solve this problem, the information available online should be verified properly before being conceived by any individual. We propose a scheme to classify the online social media posts (Tweets) with the help of the BERT (Bidirectional Encoder Representations from Transformers)-based model. Also, we compared the performance of the proposed approach with the other machine learning techniques and other State of the art techniques available. The proposed model not only classifies the tweets as relevant or irrelevant, but also creates a set of topics by which one can identify a text as relevant or irrelevant to his/her need just by just matching the keywords of the topic. To accomplish this task, after the classification of the tweets, we apply a possible topic modelling approach based on latent semantic analysis and latent Dirichlet allocation methods to identify which of the topics are mostly propagated as false information.

High-speed adaptive optics line-scan OCT for cellular-resolution optoretinography.

Optoretinography-the non-invasive, optical imaging of light-induced functional activity in the retina-stands to provide a critical biomarker for testing the safety and efficacy of new therapies as well as their rapid translation to the clinic. Optical phase change in response to light, as readily accessible in phase-resolved OCT, offers a path towards all-optical imaging of retinal function. However, typical human eye motion adversely affects phase stability. In addition, recording fast light-induced retinal events necessitates high-speed acquisition. Here, we introduce a high-speed line-scan spectral domain OCT with adaptive optics (AO), aimed at volumetric imaging and phase-resolved acquisition of retinal responses to light. By virtue of parallel acquisition of an entire retinal cross-section (B-scan) in a single high-speed camera frame, depth-resolved tomograms at speeds up to 16 kHz were achieved with high sensitivity and phase stability. To optimize spectral and spatial resolution, an anamorphic detection paradigm was introduced, enabling improved light collection efficiency and signal roll-off compared to traditional methods. The benefits in speed, resolution and sensitivity were exemplified in imaging nanometer-millisecond scale light-induced optical path length changes in cone photoreceptor outer segments. With 660 nm stimuli, individual cone responses readily segregated into three clusters, corresponding to long, middle, and short-wavelength cones. Recording such optoretinograms on spatial scales ranging from individual cones, to 100 µm-wide retinal patches offers a robust and sensitive biomarker for cone function in health and disease.

Evaluation of the Protection Provided by Hepatitis B Vaccination in India.

OBJECTIVE: In India, Hepatitis B vaccination is recommended at 6 wk except for hospital-deliveries. The authors examined protection afforded by the birth dose. METHODS: A case-control study was done. HBsAg and HBcAb were tested in 2671 children, 1 to 5 y and HBsAb was evaluated in a subset of 1413 children. Vaccination history was recorded. Cases were HBsAg carriers. In another analysis, children who got infected (HBsAg and/or HBcAb positive) were considered as cases. Exposed were the unvaccinated. In another analysis, exposed were those vaccinated without the birth dose. RESULTS: The odds ratio (OR) for HBsAg positivity with birth vaccination was 0.35 (95% CI 0.19-0.66); while with vaccination at 6 wk was 0.29 (95%CI 0.14-0.61), both compared to unvaccinated. Birth vaccination has no added protection when compared to the unvaccinated. Unvaccinated children in index study had HBsAg positivity of 4.38%. The number needed to treat (NNT) to prevent one case of HBsAg positivity was 32.6 (95% CI, 20.9 to 73.6). The odds of getting HBV infection was 0.42 (CI 0.25-0.68) with birth dose and 0.49 (CI 0.30-0.82) without the birth dose compared to the unvaccinated. Protective antibody (HBsAb) was present in about 70% of the vaccinated. In the unimmunised, in the first 2 y HBsAb protection was present in 40%. The odds ratio (OR) for HBsAb in the fully vaccinated between 4 and 5 y was 1.4 (95%CI 0.9-2.18) compared to the unvaccinated. CONCLUSIONS: The present study lends support to the pragmatic approach of the Government to vaccinate babies born at home starting at 6 wk.

Evaluation of artifact reduction in optical coherence tomography angiography with real-time tracking and motion correction technology.

Artifacts introduced by eye motion in optical coherence tomography angiography (OCTA) affect the interpretation of images and the quantification of parameters with clinical value. Eradication of such artifacts in OCTA remains a technical challenge. We developed an algorithm that recognizes five different types of motion artifacts and used it to evaluate the performance of three motion removal technologies. On en face maximum projection of flow images, the summed flow signal in each row and column and the correlation between neighboring rows and columns were calculated. Bright line artifacts were recognized by large summed flow signal. Drifts, distorted lines, and stretch artifacts exhibited abnormal correlation values. Residual lines were simultaneously a local maximum of summed flow and a local minimum of correlation. Tracking-assisted scanning integrated with motion correction technology (MCT) demonstrated higher performance than tracking or MCT alone in healthy and diabetic eyes.

Wide-field optical coherence tomography based microangiography for retinal imaging.

Optical coherence tomography angiography (OCTA) allows for the evaluation of functional retinal vascular networks without a need for contrast dyes. For sophisticated monitoring and diagnosis of retinal diseases, OCTA capable of providing wide-field and high definition images of retinal vasculature in a single image is desirable. We report OCTA with motion tracking through an auxiliary real-time line scan ophthalmoscope that is clinically feasible to image functional retinal vasculature in patients, with a coverage of more than 60 degrees of retina while still maintaining high definition and resolution. We demonstrate six illustrative cases with unprecedented details of vascular involvement in retinal diseases. In each case, OCTA yields images of the normal and diseased microvasculature at all levels of the retina, with higher resolution than observed with fluorescein angiography. Wide-field OCTA technology will be an important next step in augmenting the utility of OCT technology in clinical practice.

SWEPT SOURCE OPTICAL COHERENCE TOMOGRAPHY ANGIOGRAPHY OF NEOVASCULAR MACULAR TELANGIECTASIA TYPE 2.

BACKGROUND/PURPOSE: To image subretinal neovascularization in proliferative macular telangiectasia Type 2 (MacTel2) using swept source optical coherence tomography based microangiography (OMAG). METHODS: Patients with macular telangiectasia Type 2 were enrolled in a prospective, observational study known as the MacTel Project and evaluated using a high-speed 1,050 nm swept-source OCT prototype system. The OMAG algorithm generated en face flow images from three retinal layers, and the region bounded by the outer retina and Bruch membrane, the choriocapillaris, and the remaining choroidal vasculature. The en face OMAG images were compared with images from fluorescein angiography and indocyanine green angiography. RESULTS: Three eyes with neovascular macular telangiectasia Type 2 were imaged. The neovascularization was best identified from the en face OMAG images that included a layer between the outer retinal boundary and Bruch membrane. Optical coherence tomography based microangiography images identified these abnormal vessels better than fluorescein angiography and were comparable to the images obtained using indocyanine green angiography. In all 3 cases, OMAG identified choroidal vessels communicating with the neovascularization, and these choroidal vessels were evident in the 2 cases with indocyanine green angiography imaging. In 1 case, monthly injections of bevacizumab reduced the microvascular complexity of the neovascularization, and the telangiectatic changes within the retinal microvasculature. In another case, less frequent bevacizumab therapy was associated with growth of the subretinal neovascular complex. CONCLUSION: Optical coherence tomography based microangiography imaging provided detailed, depth-resolved information about subretinal neovascularization in macular telangiectasia Type 2 eyes demonstrating superiority to fluorescein angiography imaging, and similarities to indocyanine green angiography imaging for documenting the retinal microvascular changes, the size and extent of the neovascular complex, the communications between the neovascular complex and the choroidal circulation, and the response to monthly bevacizumab therapy.

Wide-field imaging of retinal vasculature using optical coherence tomography-based microangiography provided by motion tracking.

Optical coherence tomography (OCT)-based optical microangiography (OMAG) is a high-resolution, noninvasive imaging technique capable of providing three-dimensional in vivo blood flow visualization within microcirculatory tissue beds in the eye. Although the technique has demonstrated early clinical utility by imaging diseased eyes, its limited field of view (FOV) and the sensitivity to eye motion remain the two biggest challenges for the widespread clinical use of the technology. Here, we report the results of retinal OMAG imaging obtained from a Zeiss Cirrus 5000 spectral domain OCT system with motion tracking capability achieved by a line scan ophthalmoscope (LSO). The tracking LSO is able to guide the OCT scanning, which minimizes the effect of eye motion in the final results. We show that the tracking can effectively correct the motion artifacts and remove the discontinuities and distortions of vascular appearance due to microsaccade, leading to almost motion-free OMAG angiograms with good repeatability and reliability. Due to the robustness of the tracking LSO, we also show the montage scan protocol to provide unprecedented wide field retinal OMAG angiograms. We experimentally demonstrate a 12 x 16 mm² retinal OMAG angiogram acquired from a volunteer, which is the widest FOV retinal vasculature imaging up to now in the community.

Swept-source OCT angiography of macular telangiectasia type 2.

BACKGROUND AND OBJECTIVE: To evaluate the central macular microvascular network in patients with macular telangiectasia type 2 (MacTel2) using optical coherence tomography (OCT)-based microangiography (OMAG). PATIENTS AND METHODS: Prospective, observational study of patients with MacTel2 evaluated using a swept-source OCT (SS-OCT) prototype. OMAG was performed using a 3 mm × 3 mm central foveal raster scan. The algorithm segmented the retina into three layers. Microvascular distribution was depicted as en face images, and qualitative information was compared to fluorescein angiography (FA) images. RESULTS: OMAG detected abnormal microvasculature in all MacTel2 eyes, predominantly in the middle retinal layers with neovascularization in the outer retina. These vessels correlated well with the FA alterations. The abnormal temporal, juxtafoveal microvasculature in MacTel2 became apparent as the disease progressed and in later stages tended to extend circumferentially, with anastomotic vessels temporally. CONCLUSION: OMAG provided detailed, depth- resolved information about the perifoveal macular microvasculature in MacTel2. In most cases, images were better using OMAG than FA. The OMAG images demonstrated that most of the leakage seen on FA appeared to arise from the abnormal perifoveal microvasculature in the middle retinal layer.

Swept-source OCT angiography of the retinal vasculature using intensity differentiation-based optical microangiography algorithms.

BACKGROUND AND OBJECTIVE: To demonstrate the feasibility of using a 1,050-nm swept-source optical coherence tomography (SS-OCT) system to achieve noninvasive retinal vasculature imaging in human eyes. MATERIALS AND METHODS: Volumetric data sets were acquired using a 1-µm SS-OCT prototype that operated at a 100-kHz A-line rate. A scanning protocol designed to allow for motion contrast processing, referred to as OCT angiography or optical microangiography (OMAG), was used to scan an approximately 3 × 3­mm area in the central macular region of the retina within approximately 4.5 seconds. An intensity differentiation-based OMAG algorithm was used to extract three-dimensional retinal functional microvasculature information. RESULTS: Intensity signal differentiation generated capillary-level resolution en face OMAG images of the retina. The parafoveal capillaries were clearly visible, thereby allowing visualization of the foveal avascular zone in healthy subjects. CONCLUSION: The capability of OMAG to produce retinal vascular images was demonstrated using the 1-µm SS-OCT prototype. This technique has potential clinical value for studying retinal vasculature abnormalities.

Scalar generalized nonlinear Schrödinger equation-quantified continuum generation in an all-normal dispersion photonic crystal fiber for broadband coherent optical sources.

We quantitatively predict the observed continuum-like spectral broadening in a 90-mm weakly birefringent all-normal dispersion-flattened photonic crystal fiber pumped by 1041-nm 229-fs 76-MHz pulses from a solid-state Yb:KYW laser. The well-characterized continuum pulses span a bandwidth of up to 300 nm around the laser wavelength, allowing high spectral power density pulse shaping useful for various coherent control applications. We also identify the nonlinear polarization effect that limits the bandwidth of these continuum pulses, and therefore report the path toward a series of attractive alternative broadband coherent optical sources.

Numerical analysis of gradient index lens-based optical coherence tomography imaging probes.

We report the numerical analysis of gradient index (GRIN) lens-based optical coherence tomography imaging probes to derive optimal design parameters. Long and short working distance probes with a small focal spot are considered. In each model, the working distance and beam waist are characterized and compared for different values of length and refractive index of the probe components. We also explore the influence of the outer tubing and refractive index of the sample media. Numerical results show that the adjustment of the maximum beam diameter and focusing angle at the end of the GRIN lens surface is very important for determining the optical performance parameters of the probe.

Characterization and Analysis of Relative Intensity Noise in Broadband Optical Sources for Optical Coherence Tomography.

Relative intensity noise (RIN) is one of the most significant factors limiting the sensitivity of an optical coherence tomography (OCT) system. The existing and prevalent theory being used for estimating RIN for various light sources in OCT is questionable, and cannot be applied uniformly for different types of sources. The origin of noise in various sources differs significantly, owing to the different physical nature of photon generation. In this study, we characterize and compare RIN of several OCT light sources including superluminescent diodes (SLDs), an erbium-doped fiber amplifier, multiplexed SLDs, and a continuous-wave laser. We also report a method for reduction of RIN by amplifying the SLD light output by using a gain-saturated semiconductor optical amplifier.

Cross-correlation-based image acquisition technique for manually-scanned optical coherence tomography.

We present a novel image acquisition technique for Optical Coherence Tomography (OCT) that enables manual lateral scanning. The technique compensates for the variability in lateral scan velocity based on feedback obtained from correlation between consecutive A-scans. Results obtained from phantom samples and biological tissues demonstrate successful assembly of OCT images from manually-scanned datasets despite non-uniform scan velocity and abrupt stops encountered during data acquisition. This technique could enable the acquisition of images during manual OCT needle-guided biopsy or catheter-based imaging, and for assembly of large field-of-view images with hand-held probes during intraoperative in vivo OCT imaging.

Long-wavelength optical coherence tomography at 1.7 microm for enhanced imaging depth.

Multiple scattering in a sample presents a significant limitation to achieve meaningful structural information at deeper penetration depths in optical coherence tomography (OCT). Previous studies suggest that the spectral region around 1.7 microm may exhibit reduced scattering coefficients in biological tissues compared to the widely used wavelengths around 1.3 mum. To investigate this long-wavelength region, we developed a wavelength-swept laser at 1.7 microm wavelength and conducted OCT or optical frequency domain imaging (OFDI) for the first time in this spectral range. The constructed laser is capable of providing a wide tuning range from 1.59 to 1.75 microm over 160 nm. When the laser was operated with a reduced tuning range over 95 nm at a repetition rate of 10.9 kHz and an average output power of 12.3 mW, the OFDI imaging system exhibited a sensitivity of about 100 dB and axial and lateral resolution of 24 mum and 14 mum, respectively. We imaged several phantom and biological samples using 1.3 mum and 1.7 microm OFDI systems and found that the depth-dependent signal decay rate is substantially lower at 1.7 microm wavelength in most, if not all samples. Our results suggest that this imaging window may offer an advantage over shorter wavelengths by increasing the penetration depths as well as enhancing image contrast at deeper penetration depths where otherwise multiple scattered photons dominate over ballistic photons.

Common-path optical coherence tomography with side-viewing bare fiber probe for endoscopic optical coherence tomography.

All-fiber-optic common-path optical coherence tomography (OCT) using a side-viewing bare fiber probe has been demonstrated and analyzed. A bare single mode fiber tip is angle cleaved at approximately 49 degrees to enable side illumination due to total internal reflection. The bare fiber probe was inserted in an arterial tissue and a circumferential OCT scan was obtained. The research is aimed at realizing highly miniaturized monolithic probes for possible applications in miniature endoscopic OCT or intravascular OCT. The effects of the angle of the cleaved fiber on reference reflection and the sensitivity of the common-path OCT system have been studied theoretically. The angle cleaved fiber probe is also used in series with a microlens to analyze and optimize its performance in a common-path OCT system. Our research aims to explore the combined advantages of common-path OCT and extremely simplified miniature probe design and to discuss how it may greatly simplify the endoscopic OCT instrumentation eventually.

Fiber optic confocal laser Doppler velocimeter using an all-fiber laser source for high resolution measurements.

We demonstrate and analyze a novel fiber optic confocal laser Doppler velocimeter using an ultra-narrow linewidth all-fiber laser source centered at around 1550 nm (eye-safe region). The narrow spectral linewidth of the fiber laser (<10 kHz) is used to achieve an extremely high velocity resolution (~0.0075 m/s), which is an order of magnitude better as compared to the commonly used semiconductor diode lasers or He-Ne lasers based systems. The directional optical circulator based design used in our system is much simpler to implement and is power conserving compared to the conventional Michelson interferometer based designs. We perform Gaussian beam propagation analysis by using the ABCD law to study the performance of the confocal design. The analysis is in good accord with our experimental results. The confocal design is capable of providing ultrahigh spatial resolution (~5microm, in both lateral and longitudinal directions) for high-precision velocity distribution measurement applications.