Constructing a clinical radiographic knee osteoarthritis database using artificial intelligence tools with limited human labor: A proof of principle.

OBJECTIVE: To create a scalable and feasible retrospective consecutive knee osteoarthritis (OA) radiographic database with limited human labor using commercial and custom-built artificial intelligence (AI) tools. METHODS: We applied four AI tools, two commercially available and two custom-built tools, to analyze 6 years of clinical consecutive knee radiographs from patients aged 35-79 at the University of Copenhagen Hospital, Bispebjerg-Frederiksberg Hospital, Denmark. The tools provided Kellgren-Lawrence (KL) grades, joint space widths, patella osteophyte detection, radiographic view detection, knee joint implant detection, and radiographic marker detection. RESULTS: In total, 25,778 knee radiographs from 8575 patients were included in the database after excluding inapplicable radiographs, and 92.5% of the knees had a complete OA dataset. Using the four AI tools, we saved about 800 hours of radiologist reading time and only manually reviewed 16.0% of the images in the database. CONCLUSIONS: This study shows that clinical knee OA databases can be built using AI with limited human reading time for uniform grading and measurements. The concept is scalable temporally and across geographic regions and could help diversify further OA research by efficiently including radiographic knee OA data from different populations globally. We can prevent data dredging and overfitting OA theories on existing trite cohorts by including various gene pools and continuous expansion of new clinical cohorts. Furthermore, the suggested tools and applied approaches provide an ability to retest previous hypotheses and test new hypotheses on real-life clinical data with current disease prevalence and trends.

Hypoglycemia-Associated EEG Changes in Prepubertal Children With Type 1 Diabetes.

BACKGROUND: The purpose of this study was to explore the possible difference in the electroencephalogram (EEG) pattern between euglycemia and hypoglycemia in children with type 1 diabetes (T1D) during daytime and during sleep. The aim is to develop a hypoglycemia alarm based on continuous EEG measurement and real-time signal processing. METHOD: Eight T1D patients aged 6-12 years were included. A hyperinsulinemic hypoglycemic clamp was performed to induce hypoglycemia both during daytime and during sleep. Continuous EEG monitoring was performed. For each patient, quantitative EEG (qEEG) measures were calculated. A within-patient analysis was conducted comparing hypoglycemia versus euglycemia changes in the qEEG. The nonparametric Wilcoxon signed rank test was performed. A real-time analyzing algorithm developed for adults was applied. RESULTS: The qEEG showed significant differences in specific bands comparing hypoglycemia to euglycemia both during daytime and during sleep. In daytime the EEG-based algorithm identified hypoglycemia in all children on average at a blood glucose (BG) level of 2.5 ± 0.5 mmol/l and 18.4 (ranging from 0 to 55) minutes prior to blood glucose nadir. During sleep the nighttime algorithm did not perform. CONCLUSIONS: We found significant differences in the qEEG in euglycemia and hypoglycemia both during daytime and during sleep. The algorithm developed for adults detected hypoglycemia in all children during daytime. The algorithm had too many false alarms during the night because it was more sensitive to deep sleep EEG patterns than hypoglycemia-related EEG changes. An algorithm for nighttime EEG is needed for accurate detection of nocturnal hypoglycemic episodes in children. This study indicates that a hypoglycemia alarm may be developed using real-time continuous EEG monitoring.

A Novel Noninvasive Method for Measuring Fatigability of the Quadriceps Muscle in Noncooperating Healthy Subjects.

BACKGROUND: Critical illness is associated with muscle weakness leading to long-term functional limitations. OBJECTIVES: To assess the reliability of a novel method for evaluating fatigability of the quadriceps muscle in noncooperating healthy subjects. METHODS: On two occasions, separated by seven days, nonvoluntary isometric contractions (twitch and tetanic) of the quadriceps femoris muscle evoked by transcutaneous electrical muscle stimulation were recorded in twelve healthy adults. For tetanic contractions, the Fatigue Index (ratio of peak torque values) and the slope of the regression line of peak torque values were primary outcome measures. For twitch contractions, maximum peak torque and rise time were calculated. Relative (intraclass correlation, ICC3.1) and absolute (standard error of measurement, SEM) reliability were assessed and minimum detectable change was calculated using a 95% confidence interval (MDC95%). RESULTS: The Fatigue Index (ICC3.1, 0.84; MDC95%, 0.12) and the slope of the regression line (ICC3.1, 0.99; MDC95%, 0.03) showed substantial relative and absolute reliability during the first 15 and 30 contractions, respectively. CONCLUSION: This method for assessing fatigability of the quadriceps muscle produces reliable results in healthy subjects and may provide valuable data on quantitative changes in muscle working capacity and treatment effects in patients who are incapable of producing voluntary muscle contractions.

Effects of training and weight support on muscle activation in Parkinson's disease.

The aim of this study was to investigate the effect of high-intensity locomotor training on knee extensor and flexor muscle activation and adaptability to increased body-weight (BW) support during walking in patients with Parkinson's disease (PD). Thirteen male patients with idiopathic PD and eight healthy participants were included. The PD patients completed an 8-week training program on a lower-body, positive-pressure treadmill. Knee extensor and flexor muscles activation during steady treadmill walking (3 km/h) were measured before, at the mid-point, and after training. Increasing BW support decreased knee extensor muscle activation (normalization) and increased knee flexor muscle activation (abnormal) in PD patients when compared to healthy participants. Training improved flexor peak muscle activation adaptability to increased (BW) support during walking in PD patients. During walking without BW support shorter knee extensor muscle off-activation time and increased relative peak muscle activation was observed in PD patients and did not improve with 8 weeks of training. In conclusion, patients with PD walked with excessive activation of the knee extensor and flexor muscles when compared to healthy participants. Specialized locomotor training may facilitate adaptive processes related to motor control of walking in PD patients.

Biomechanical and nonfunctional assessment of physical capacity in male ICU survivors.

OBJECTIVES: ICU admission is associated with decreased physical function for years after discharge. The underlying mechanisms responsible for this muscle function impairment are undescribed. The aim of this study was to describe the biomechanical properties of the quadriceps muscle in ICU survivors 12 months after ICU discharge. DESIGN: Case-control study with consecutive inclusion of ICU survivors and age- and sex-matched controls. SETTING: Patients were treated at a mixed 18-bed ICU at a tertiary care university hospital and tested at a biomechanical university laboratory. PATIENTS: We included 16 male ICU patients (Acute Physiology and Chronic Health Evaluation II score 20 ± 7, mean ± SD), who had stayed in the ICU >72 hrs and survived to 12 months and 15 age- and sex-matched controls. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: An extensive battery of biomechanical tests, including maximum, fast, and endurance contractions, was administered during isometric knee extensions while simultaneously recording surface electromyography (quadriceps and hamstrings). Compared to controls, ICU survivors had reduced maximal voluntary torque (22%, 179 ± 64 Nm vs. 230 ± 57 Nm, p = 0.03), absolute rate of force development (50%, 868 ± 372 Nm/sec vs. 1739 ± 470 Nm/sec, p < 0.001) and relative rate of force development (32%, 512 ± 260% maximum voluntary contraction/sec vs. 754 ± 189% maximum voluntary contraction/sec, p < 0.01), and endurance time (40%, 136 ± 84 sec vs. 226 ± 111 sec, p < 0.02). Rate of force development, but not maximal voluntary torque, was significantly reduced after adjusting for muscle mass. Electromyography data indicated no impairment of motor activation strategy or central motor drive. Also, no difference in reaction time was found between patients and controls. CONCLUSIONS: ICU survivors had reduced rate of force development and muscular endurance 1 yr after ICU discharge. Our data indicate that the functional deficits experienced by ICU survivors originate in muscle tissue rather than the nervous system. Also, increased attention to velocity-orientated exercise during rehabilitation of ICU patients may have the potential to better physical outcome after critical illness.

Tremor irregularity, torque steadiness and rate of force development in Parkinson's disease.

We investigated lower-extremity isometric tremor Approximate Entropy (irregularity), torque steadiness and rate of force development (RFD) and their associations to muscle activation strategy during isometric knee extensions in patients with Parkinson's disease (PD). Thirteen male patients with idiopathic PD and 15 neurologically healthy matched controls performed isometric maximal contractions (extension/flexion) as well as steady submaximal and powerful isometric knee extensions. The patients with PD showed decreased isometric tremor irregularity. Torque steadiness was reduced in PD and the patients had increased muscle coactivation. A markedly lower RFD was found in PD and the decreased RFD correlated with reduced agonist muscle activation. Furthermore, patient RFD correlated with the Movement-Disorder-Society-Unified-Parkinson's-Disease-Rating-Scale 3 (motor part) scores. We concluded that both knee isometric tremor Approximate Entropy and torque steadiness clearly differentiate between patients with PD and healthy controls. Furthermore, severely compromised RFD was found in patients with PD and was associated with decreased agonist muscle activation.

Improved clinical status, quality of life, and walking capacity in Parkinson's disease after body weight-supported high-intensity locomotor training.

OBJECTIVE: To evaluate the effect of body weight-supported progressive high-intensity locomotor training in Parkinson's disease (PD) on (1) clinical status; (2) quality of life; and (3) gait capacity. DESIGN: Open-label, fixed sequence crossover study. SETTING: University motor control laboratory. PARTICIPANTS: Patients (N=13) with idiopathic PD (Hoehn and Yahr stage 2 or 3) and stable medication use. INTERVENTIONS: Patients completed an 8-week (3 × 1h/wk) training program on a lower-body positive-pressure treadmill. Body weight support was used to facilitate increased intensity and motor challenges during treadmill training. The training program contained combinations of (1) running and walking intervals, (2) the use of sudden changes (eg, in body weight support and speed), (3) different types of locomotion (eg, chassé, skipping, and jumps), and (4) sprints at 50 percent body weight. MAIN OUTCOME MEASURES: The Movement Disorders Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS), Parkinson's Disease Questionnaire-39 items (PDQ-39), and the six-minute walk test were conducted 8 weeks before and pre- and posttraining. RESULTS: At the end of training, statistically significant improvements were found in all outcome measures compared with the control period. Total MDS-UPDRS score changed from (mean ± 1SD) 58±18 to 47±18, MDS-UPDRS motor part score changed from 35±10 to 29±12, PDQ-39 summary index score changed from 22±13 to 13±12, and the six-minute walking distance changed from 576±93 to 637±90m. CONCLUSIONS: Body weight-supported progressive high-intensity locomotor training is feasible and well tolerated by patients with PD. The training improved clinical status, quality of life, and gait capacity significantly.

Ankle torque steadiness is related to muscle activation variability and coactivation in children with cerebral palsy.

The aims of this study were to: (1) investigate the significance of muscle activation variability and coactivation for the ability to perform steady submaximal ankle torque (torque steadiness) in healthy children and those with cerebral palsy (CP), and (2) assess ankle function during isometric contractions in those children. Fourteen children with CP who walked with equinus foot deformity and 14 healthy (control) children performed maximal and steady submaximal ankle dorsi- and plantarflexions. Dorsiflexion torque steadiness was related to agonist and antagonist muscle activation variability as well as the plantarflexor coactivation level in children with CP (r > 0.624, P < 0.03). Moreover, children with CP displayed reduced maximal torque and submaximal torque steadiness of both dorsi- and plantarflexion compared with controls (P < 0.05). Both muscle groups may benefit from strength training, as they exhibit poor submaximal control and weakness in children with CP.