Game based assessments of cognitive ability in recruitment: Validity, fairness and test-taking experience

Gamification and machine learning are emergent technologies in recruitment, promising to improve the user experience and fairness of assessments. We test this by validating a game based assessment of cognitive ability with a machine learning based scoring algorithm optimised for validity and fairness. We use applied data from 11,574 assessment completions. The assessment has convergent validity (r = 0.5) and test-retest reliability (r = 0.68). It maintains fairness in a separate sample of 3,107 job applicants, showing that fairness-optimised machine learning can improve outcome parity issues with cognitive ability tests in recruitment settings. We show that there are no significant gender differences in test taking anxiety resulting from the games, and that anxiety does not directly predict game performance, supporting the notion that game based assessments help with test taking anxiety. Interactions between anxiety, gender and performance are explored. Feedback from 4,778 job applicants reveals a Net Promoter score of 58, indicating more applicants support than dislike the assessment, and that games deliver a positive applicant experience in practise. Satisfaction with the format is high, but applicants raise face validity concerns over the abstract games. We encourage the use of gamification and machine learning to improve the fairness and user experience of psychometric tests.

Oxidative Stress and Cardiovascular Risk in Type 1 Diabetes Mellitus: Insights From the DCCT/EDIC Study

Background

Hyperglycemia leading to increased oxidative stress is implicated in the increased risk for the development of macrovascular and microvascular complications in patients with type 1 diabetes mellitus.

Methods and Results

A random subcohort of 349 participants was selected from the DCCT / EDIC (Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications) cohort. This included 320 controls and 29 cardiovascular disease cases that were augmented with 98 additional known cases to yield a case cohort of 447 participants (320 controls, 127 cases). Biosamples from DCCT baseline, year 1, and closeout of DCCT , and 1 to 2 years post‐ DCCT ( EDIC years 1 and 2) were measured for markers of oxidative stress, including plasma myeloperoxidase, paraoxonase activity, urinary F 2α isoprostanes, and its metabolite, 2,3 dinor‐8 iso prostaglandin F 2α . Following adjustment for glycated hemoblobin and weighting the observations inversely proportional to the sampling selection probabilities, higher paraoxonase activity, reflective of antioxidant activity, and 2,3 dinor‐8 iso prostaglandin F 2α , an oxidative marker, were significantly associated with lower risk of cardiovascular disease (−4.5% risk for 10% higher paraoxonase, P <0.003; −5.3% risk for 10% higher 2,3 dinor‐8 iso prostaglandin F 2α , P =0.0092). In contrast, the oxidative markers myeloperoxidase and F 2α isoprostanes were not significantly associated with cardiovascular disease after adjustment for glycated hemoblobin. There were no significant differences between DCCT intensive and conventional treatment groups in the change in all biomarkers across time segments.

Conclusions

Heightened antioxidant activity (rather than diminished oxidative stress markers) is associated with lower cardiovascular disease risk in type 1 diabetes mellitus, but these biomarkers did not change over time with intensification of glycemic control.

Clinical Trial Registration

URL : https://www.clinicaltrials.gov . Unique identifiers: NCT 00360815 and NCT 00360893.

Computing with networks of spiking neurons on a biophysically motivated floating-gate based neuromorphic integrated circuit

Results are presented from several spiking network experiments performed on a novel neuromorphic integrated circuit. The networks are discussed in terms of their computational significance, which includes applications such as arbitrary spatiotemporal pattern generation and recognition, winner-take-all competition, stable generation of rhythmic outputs, and volatile memory. Analogies to the behavior of real biological neural systems are also noted. The alternatives for implementing the same computations are discussed and compared from a computational efficiency standpoint, with the conclusion that implementing neural networks on neuromorphic hardware is significantly more power efficient than numerical integration of model equations on traditional digital hardware.

A learning-enabled neuron array IC based upon transistor channel models of biological phenomena

We present a single-chip array of 100 biologically-based electronic neuron models interconnected to each other and the outside environment through 30,000 synapses. The chip was fabricated in a standard 350 nm CMOS IC process. Our approach used dense circuit models of synaptic behavior, including biological computation and learning, as well as transistor channel models. We use Address-Event Representation (AER) spike communication for inputs and outputs to this IC. We present the IC architecture and infrastructure, including IC chip, configuration tools, and testing platform. We present measurement of small network of neurons, measurement of STDP neuron dynamics, and measurement from a compiled spiking neuron WTA topology, all compiled into this IC.

Neural dynamics in reconfigurable silicon

A neuromorphic analog chip is presented that is capable of implementing massively parallel neural computations while retaining the programmability of digital systems. We show measurements from neurons with Hopf bifurcations and integrate and fire neurons, excitatory and inhibitory synapses, passive dendrite cables, coupled spiking neurons, and central pattern generators implemented on the chip. This chip provides a platform for not only simulating detailed neuron dynamics but also uses the same to interface with actual cells in applications such as a dynamic clamp. There are 28 computational analog blocks (CAB), each consisting of ion channels with tunable parameters, synapses, winner-take-all elements, current sources, transconductance amplifiers, and capacitors. There are four other CABs which have programmable bias generators. The programmability is achieved using floating gate transistors with on-chip programming control. The switch matrix for interconnecting the components in CABs also consists of floating-gate transistors. Emphasis is placed on replicating the detailed dynamics of computational neural models. Massive computational area efficiency is obtained by using the reconfigurable interconnect as synaptic weights, resulting in more than 50 000 possible 9-b accurate synapses in 9 mm(2).

Gap junctional communication in human osteoclasts in vitro and in vivo

Bone-forming cells are known to be coupled by gap junctions, formed primarily by connexin43 (Cx43). The role of Cx43 in osteoclasts has so far only been studied in rodents, where Cx43 is important for fusion of mononuclear precursors to osteoclasts. Given the potential importance for human diseases with pathologically altered osteoclasts, we asked whether a similar influence of Cx43 can also be observed in osteoclasts of human origin. For this purpose, Cx43 mRNA expression was studied in a time course experiment of human osteoclast differentiation by RT-PCR. Localization of Cx43 in these cells was determined by immunohistochemistry and confocal microscopy. For the assessment of the effect of gap junction inhibition on cell fusion, gap junctions were blocked with heptanol during differentiation of the cells and the cells were then evaluated for multinuclearity. Paraffin sections of healthy bone and bone from patients with Paget's disease and giant cell tumour of the bone were used to study Cx43 expression in vivo. We found mRNA and protein expression of Cx43 in fully differentiated osteoclasts as well as in precursor cells. This expression decreased in the course of differentiation. Consistently, we found a lower expression of Cx43 in osteoclasts than in bone marrow precursor cells in the histology of healthy human bone. Blockade of gap junctional communication by heptanol led to a dose-dependent decrease in multinuclearity, suggesting that gap junctional communication precedes cell fusion of human osteoclasts. Indeed, we found a particularly strong expression of Cx43 in the giant osteoclasts of patients with Paget's disease and giant cell tumour of the bone. These results show that gap junctional communication is important for fusion of human mononuclear precursor cells to osteoclasts and that gap junctional Cx43 might play a role in the regulation of size and multinuclearity of human osteoclasts in vivo.

Sick day management using blood 3-hydroxybutyrate (3-OHB) compared with urine ketone monitoring reduces hospital visits in young people with T1DM: a randomized clinical trial

AIMS

Diabetic ketoacidosis (DKA), a life-threatening acute complication of Type 1 diabetes, may be preventable with frequent monitoring of glycaemia and ketosis along with timely supplemental insulin. This prospective, two-centre study assessed sick day management using blood 3-hydroxybutyrate (3-OHB) monitoring compared with traditional urine ketone testing, aimed at averting emergency assessment and hospitalization.

METHODS

One hundred and twenty-three children, adolescents and young adults, aged 3-22 years, and their families received sick day education. Participants were randomized to receive either a blood glucose monitor that also measures blood 3-OHB (blood ketone group, n = 62) or a monitor plus urine ketone strips (urine ketone group, n = 61). All were encouraged to check glucose levels > or = 3 times daily and to check ketones during acute illness or stress, when glucose levels were consistently elevated (> or = 13.9 mmol/l on two consecutive readings), or when symptoms of DKA were present. Frequency of sick days, hyperglycaemia, ketosis, and hospitalization/emergency assessment were ascertained prospectively for 6 months.

RESULTS

There were 578 sick days during 21,548 days of follow-up. Participants in the blood ketone group checked ketones significantly more during sick days (276 of 304 episodes, 90.8%) than participants in the urine ketone group (168 of 274 episodes, 61.3%) (P < 0.001). The incidence of hospitalization/emergency assessment was significantly lower in the blood ketone group (38/100 patient-years) compared with the urine ketone group (75/100 patient-years) (P = 0.05).

CONCLUSIONS

Blood ketone monitoring during sick days appears acceptable to and preferred by young people with Type 1 diabetes. Routine implementation of blood 3-OHB monitoring for the management of sick days and impending DKA can potentially reduce hospitalization/emergency assessment compared with urine ketone testing and offers potential cost savings.

Insulin lispro lowers postprandial glucose in prepubertal children with diabetes

OBJECTIVE

This study compared the glucose-lowering effect of insulin lispro, given before or after meals, with regular human insulin given before meals in prepubertal children with diabetes.

RESEARCH DESIGN AND METHODS

A 3-way crossover, open-label study involving 61 prepubertal children (ages 2.9-11.4 years) with type 1 diabetes. The children were randomly assigned to receive regular human insulin 30 to 45 minutes before meals, insulin lispro within 15 minutes before or immediately after meals, combined with basal insulin. Each treatment lasted 3 months. Hemoglobin A(1c) levels and home glucose monitoring profiles were measured at the end of each treatment period.

RESULTS

Treatment with insulin lispro before breakfast resulted in lower 2-hour postprandial glucose values than regular human insulin (11.7 +/- 4.4 mmol/L vs 15.0 +/- 5.4 mmol/L). Similarly, insulin lispro given before dinner resulted in lower blood glucose values 2 hours postprandially (8.8 +/- 5.0 mmol/L vs 10.8 +/- 5.4 mmol/L) than regular human insulin. When insulin lispro was administered after meals, the 2-hour glucose levels were between those seen with either insulin lispro or regular human insulin given before meals. The number and types of adverse events, the rates of hypoglycemia, and the HbA(1c) levels did not differ among the 3 therapies.

CONCLUSIONS

In prepubertal children, insulin lispro given before meals is safe and significantly lowers postprandial glucose levels after breakfast and dinner compared with regular human insulin, and insulin lispro given after the meal provides similar benefits as regular human insulin before the meal.

Targeting of thylakoid proteins by the delta pH-driven twin-arginine translocation pathway requires a specific signal in the hydrophobic domain in conjunction with the twin-arginine motif

Superficially similar cleavable targeting signals specify whether lumenal proteins are transported across the thylakoid membrane by a Sec- or delta pH-dependent pathway. A twin-arginine motif is essential but not sufficient to direct delta pH-dependent targeting; here we show that a second determinant is located in the hydrophobic region. A highly hydrophobic amino acid is found either two or three residues C-terminal to the twin-arginine in all known transfer peptides for the delta pH-dependent system, and substitution of this residue in the 23-kDa (23K) peptide markedly inhibits translocation. Further, whereas the insertion of twin-arginine in a Sec-dependent precursor does not permit efficient delta pH-dependent targeting, the simultaneous presence of a leucine at the +3 position (relative to the RR) enables the peptide to function as efficiently as an authentic transfer peptide. RRNVL, RRAAL and RRALA within a Sec targeting signal all support efficient delta pH-dependent targeting, RRNVA is less effective and RRNAA/RRNAG are totally ineffective. We conclude that the core signal for this pathway is a twin-arginine together with an adjacent hydrophobic determinant.

Pathway specificity for a delta pH-dependent precursor thylakoid lumen protein is governed by a 'Sec-avoidance' motif in the transfer peptide and a 'Sec-incompatible' mature protein

Cleavable N-terminal targeting signals direct the translocation of lumenal proteins across the chloroplast thylakoid membrane by either a Sec-type or delta pH-driven protein translocase. The targeting signals specify choice of translocation pathway, yet all resemble typical bacterial 'signal' peptides in possessing a charged N-terminus (N-domain), hydrophobic core region (H-domain) and more polar C-terminal region (C-domain). We have previously shown that a twin-arginine motif in the N-domain is essential for targeting by the delta pH-dependent pathway, but it has remained unclear why targeting signals for this system (transfer peptides) are not recognized by the Sec apparatus. We show here that the conserved charge distribution around the H-domain in the 23K transfer peptide (twin-Arg in the N-domain, Lys in the C-domain) constitutes a 'Sec-avoidance' signal. The C-domain Lys, while not important for delta pH-dependent targeting, is the only barrier to Sec-dependent translocation; its removal generates an apparently perfect signal peptide. Conversely, insertion of twin-Arg into the N-domain of a Sec substrate has little effect, as has insertion of a C-domain Lys, but the combined substitutions almost totally block transport. We also show that the 23K mature protein is incapable of being targeted by the Sec pathway, and it is proposed that the role of the Sec-avoidance motif in the transfer peptide is to prevent futile interactions with the Sec apparatus.

Unusual characteristics of amino-terminal and hydrophobic domains in nuclear-encoded thylakoid signal peptides

Thylakoid transfer signals carry information specifying translocation by either a Sec- or delta pH-dependent protein translocator in the chloroplast thylakoid membrane, yet all resemble classical signal peptides in overall structural terms. Comparison of known transfer signals reveals two differences: (a) signals for the delta pH-driven system invariably contain a critical twin-arginine (Arg-Arg) motif prior to the hydrophobic (H) domain, whereas known Sec-dependent signals contain lysine, and (b) the H-domains of Sec-dependent signals are generally longer. Previous work has shown that a twin-Arg motif before the H-domain is critical for targeting by the delta pH-dependent pathway; in this report we show that the charge characteristics of this region are not important for sorting by the Sec pathway. Twin-Lys, twin-Arg or single Arg are all acceptable to the Sec system, although single Lys/Arg is preferred. The single Lys in pre-plastocyanin can even be replaced by an uncharged residue without apparent effect. We have also generated a pre-plastocyanin mutant containing an H-domain which, in terms of hydropathy profile, is identical to that of a delta pH-dependent protein. This mutant is also transported efficiently by the Sec system, demonstrating that hydrophobicity per se is not a key sorting determinant. However, the characteristics of the H-domain may be important in avoiding a different form of mis-targeting: to the endoplasmic reticulum. Thylakoid signal peptides have undergone substantial structural changes during the evolution of the chloroplast from endosymbiotic cyanobacterium: plastid-encoded and cyanobacterial signals contain H-domains that are highly hydrophobic and enriched in Leu and aromatic residues, whereas nuclear-encoded counterparts are Ala-rich and far less hydrophobic. We speculate that this trend may reflect a need to avoid mistargeting through recognition by cytosolic signal recognition particle, which preferentially interacts with more hydrophobic signal peptides.