Research on different machine learning (ML) has become incredibly popular during the past few decades. However, for some researchers not familiar with statistics, it might be difficult to understand how to evaluate the performance of ML models and compare them with each other. Here, we introduce the most common evaluation metrics used for the typical supervised ML tasks including binary, multi-class, and multi-label classification, regression, image segmentation, object detection, and information retrieval. We explain how to choose a suitable statistical test for comparing models, how to obtain enough values of the metric for testing, and how to perform the test and interpret its results. We also present a few practical examples about comparing convolutional neural networks used to classify X-rays with different lung infections and detect cancer tumors in positron emission tomography images.

Purpose Brown adipose tissue (BAT) is considered a potential target for combatting obesity, as it produces heat instead of ATP in cellular respiration due to uncoupling protein-1 (UCP-1) in mitochondria. However, BAT-specific thermogenic capacity, in comparison to whole-body thermogenesis during cold stimulus, is still controversial. In our present study, we aimed to determine human BAT oxygen consumption with [¹⁵O]O₂ positron emission tomography (PET) imaging. Further, we explored whether BAT-specific energy expenditure (EE) is associated with BAT blood flow, non-esterified fatty acid (NEFA) uptake, and whole-body EE. Methods Seven healthy study subjects were studied at two different scanning sessions, 1) at room temperature (RT) and 2) with acute cold exposure. Radiotracers [¹⁵O]O₂, [¹⁵O]H₂O, and [¹⁸F]FTHA were given for the measurements of BAT oxygen consumption, blood flow, and NEFA uptake, respectively, with PET-CT. Indirect calorimetry was performed to assess differences in whole-body EE between RT and cold. Results BAT-specific EE and oxygen consumption was higher during cold stimulus (approx. 50 %); similarly, whole-body EE was higher during cold stimulus (range 2–47 %). However, there was no association in BAT-specific EE and whole-body EE. BAT-specific EE was found to be a minor contributor in cold induced whole-body thermogenesis (almost 1 % of total whole-body elevation in EE). Certain deep muscles in the cervico-thoracic region made a major contribution to this cold-induced thermogenesis (CIT) without any visual signs or individual perception of shivering. Moreover, BAT-specific EE associated with BAT blood flow and NEFA uptake both at RT and during cold stimulus. Conclusion Our study suggests that BAT is a minor and deep muscles are a major contributor to CIT. In BAT, both in RT and during cold, cellular respiration is linked with circulatory NEFA uptake.

Carimas is a multi-purpose medical imaging data processing tool, which can be used to visualize, analyze, and model different medical images in research. Originally, it was developed only for positron emission tomography data in 2009, but the use of this software has extended to many other tomography imaging modalities, such as computed tomography and magnetic resonance imaging. Carimas is especially well-suited for analysis of three- and four-dimensional image data and creating polar maps in modeling of cardiac perfusion. This article explores various parts of Carimas, including its key features, program structure, and application possibilities.

Objective To investigate brain amyloid pathology in a dementia-risk population defined as cardiovascular risk factors, aging, and dementia risk (CAIDE) score of at least 6 but with normal cognition and to examine associations between brain amyloid load and cognitive performance and vascular risk factors. Methods A subgroup of 48 individuals from the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) main study participated in brain ¹¹C-Pittsburgh compound B (PiB)-PET imaging, brain MRI, and neuropsychological assessment at the beginning of the study. Lifestyle/vascular risk factors were determined as body mass index, blood pressure, total and low-density lipoprotein cholesterol, and glucose homeostasis model assessment. White matter lesions were visually rated from MRIs by a semiquantitative Fazekas score. Results Twenty participants (42%) had a positive PiB-PET on visual analysis. The PiB-positive group performed worse in executive functioning tests, included more participants with APOE ε4 allele (50%), and showed slightly better glucose homeostasis compared to PiB-negative participants. PiB-positive and -negative participants did not differ significantly in other cognitive domain scores or other vascular risk factors. There was no significant difference in Fazekas score between the PiB groups. Conclusions The high percentage of PiB-positive participants provides evidence of a successful recruitment process of the at-risk population in the main FINGER intervention trial. The results suggest a possible association between early brain amyloid accumulation and decline in executive functions. APOE ε4 was clearly associated with amyloid positivity, but no other risk factor was found to be associated with positive PiB-PET.

Objective: Meal ingestion is followed by a redistribution of blood flow (BF) within the splanchnic region contributing to nutrient absorption, insulin secretion and glucose disposal, but factors regulating this phenomenon in humans are poorly known. The aim of the present study was to evaluate the organ-specific changes in BF during a mixed-meal and incretin infusions. Design: A non-randomized intervention study of 10 healthy adults to study splanchnic BF regulation was performed. Methods: Effects of glucose-dependent insulinotrophic polypeptide (GIP) and glucagon-like peptide 1 (GLP-1) infusions and mixed-meal were tested in 10 healthy, glucose tolerant subjects using PET-MRI multimodal imaging technology. Intestinal and pancreatic BF and blood volume (BV) were measured with ¹⁵O-water and ¹⁵O-carbon monoxide, respectively. Results: Ingestion of a mixed-meal led to an increase in pancreatic and jejunal BF, whereas duodenal BF was unchanged. Infusion of GIP and GLP-1 reduced BF in the pancreas. However, GIP infusion doubled blood flow in the jejunum with no effect of GLP-1. Conclusion: Together, our data suggest that meal ingestion leads to increases in pancreatic BF accompanied by a GIP-mediated increase in jejunal but not duodenal blood flow.