Quantifying and Improving the Performance of Speech Recognition Systems on Dysphonic Speech.

OBJECTIVE: This study seeks to quantify how current speech recognition systems perform on dysphonic input and if they can be improved. STUDY DESIGN: Experimental machine learning methods based on a retrospective database. SETTING: Single academic voice center. METHODS: A database of dysphonic speech recordings was created and tested against 3 speech recognition platforms. Platform performance on dysphonic voice input was compared to platform performance on normal voice input. A custom speech recognition model was trained on voice from patients with spasmodic dysphonia or vocal cord paralysis. Custom model performance was compared to base model performance. RESULTS: All platforms performed well on normal voice, and 2 platforms performed significantly worse on dysphonic speech. Accuracy metrics on dysphonic speech returned values of 84.55%, 88.57%, and 93.56% for International Business Machines (IBM) Watson, Amazon Transcribe, and Microsoft Azure, respectively. The secondary analysis demonstrated that the lower performance of IBM Watson and Amazon Transcribe was driven by performance on spasmodic dysphonia and vocal fold paralysis. Thus, a custom model was built to increase the accuracy of these pathologies on the Microsoft platform. Overall, the performance of the custom model on dysphonic voices was 96.43% and on normal voices was 97.62%. CONCLUSION: Current speech recognition systems generally perform worse on dysphonic speech than on normal speech. We theorize that poor performance is a consequence of a lack of dysphonic voices in each platform's original training dataset. We address this limitation with transfer learning used to increase the performance of these systems on all dysphonic speech.

Effects of adolescent methylphenidate administration on methamphetamine conditioned place preference in an animal model of attention-deficit/hyperactivity disorder: Examination of potential sex differences.

BACKGROUND: Individuals with attention-deficit/hyperactivity disorder (ADHD) are more likely to be diagnosed with a substance use disorder; however, the effects of long-term psychostimulant treatment on addiction are mixed. Preclinical studies are useful for further elucidating the relationship between ADHD and addiction-like behaviors, but these studies have focused on male subjects only. The goal of the current study was to determine if early-life administration of methylphenidate (MPH) augments methamphetamine (METH) conditioned place preference (CPP) and/or potentiates reinstatement of CPP in both male and female rats. METHODS: Male and female spontaneously hypertensive rats (SHRs) and Wistar-Kyoto rats (WKYs) received either MPH (1.5mg/kg; p.o.) or vehicle (1.0ml/kg) during adolescence (postnatal day [PND] ~29-57). Two weeks after cessation of MPH treatment, rats were tested for METH CPP (1.0mg/kg or 2.0mg/kg; s.c.). Rats were then given extinction sessions. Once rats met extinction criteria, they were tested for reinstatement of CPP following a priming injection of METH (0.25mg/kg; s.c.). RESULTS: All groups developed METH CPP, except vehicle-treated SHR males and vehicle-treated WKY females conditioned with the higher dose of METH (2.0mg/kg). Female SHRs treated with MPH showed greater reinstatement of METH CPP compared to female SHRs treated with vehicle. Adolescent MPH treatment did not augment the locomotor-stimulant effects of METH in adulthood. CONCLUSIONS: These results demonstrate the importance of considering biological sex when prescribing psychostimulant medications for ADHD as long-term MPH administration may increase the risk of continued drug use in females with ADHD following a period of abstinence.

Effects of NMDA receptor antagonists on behavioral economic indices of cocaine self-administration.

BACKGROUND: Currently, there are no FDA-approved medications for the treatment of psychostimulant (e.g., cocaine) use disorders. Because the GluN2B subunit of the glutamate N-methyl-D-aspartate (NMDA) receptor is an important mediator of addiction-like behaviors, the goal of the current study was to determine if the GluN2B-selective antagonist Ro 63-1908 is efficacious in attenuating cocaine self-administration. METHODS: Adult Sprague Dawley rats (24 males and 11 females) were implanted with indwelling catheters and were trained to self-administer cocaine (0.75 mg/kg/inf). Rats were then trained in a threshold procedure, in which the dose of cocaine decreased across six 6-min blocks (0.75, 0.27, 0.08, 0.03, 0.01, 0.003 mg/kg/inf). This procedure allowed for the quantification of behavioral economic indices of drug self-administration. Following training in the threshold procedure, rats were treated with the GluN2B-selective antagonist Ro 63-1908 (0, 0.1, 0.3, 1.0 mg/kg; s.c.). Rats also received treatments of the NMDA receptor channel blocker MK-801 (0, 0.01, 0.03, 0.06 mg/kg; s.c.). RESULTS: Blocking NMDA receptors decreased initial intake (i.e., consumption during the first block), although Ro 63-1908 and MK-801 increased area under the curve (global measure of demand) and decreased demand elasticity, an effect observed primarily in males. Neither drug affected demand intensity (i.e., consumption of cocaine at a minimally constrained price). CONCLUSIONS: While blocking the NMDA receptor decreases initial intake of cocaine, NMDA receptor antagonists make cocaine more inelastic with increasing price. These results suggest that NMDA receptor antagonists can exacerbate addiction-like behaviors during self-administration during extinction-like conditions that are observed in later blocks of the threshold procedure.

Effects of the GluN2B-selective antagonist Ro 63-1908 on acquisition and expression of methamphetamine conditioned place preference in male and female rats.

BACKGROUND: Methamphetamine abuse has increased significantly in recent years. Currently, there are no FDA-approved pharmacotherapies for the treatment of methamphetamine use disorder. The goal of the current study was to determine if the N-methyl-d-aspartate (NMDA) GluN2B-selective antagonist Ro 63-1908 can block the conditioned rewarding effects of methamphetamine as assessed in conditioned place preference (CPP). METHODS: Two main experiments were conducted. In the first experiment, male (n = 24) and female (n = 24) rats received either vehicle or Ro 63-1908 (1.0-10.0 mg/kg) 30 min prior to the posttest to determine if blocking the GluN2B subunit attenuates expression of methamphetamine CPP. In the second experiment, male (n = 18) and female (n = 18) rats received either vehicle or Ro 63-1908 (1.0 or 3.0 mg/kg) 30 min prior to each conditioning session to determine if blocking the GluN2B subunit attenuates acquisition of methamphetamine CPP. RESULTS: Ro 63-1908 (3.0 mg/kg) blocked acquisition of methamphetamine CPP in male rats, but only attenuated CPP in female rats. Ro 63-1908 did not alter expression of CPP in either sex. Increasing the dose of Ro 63-1908 (10.0 mg/kg) failed to block acquisition of CPP in an additional group of female rats (n = 6). A control experiment showed that Ro 63-1908 (3.0 mg/kg) did not produce CPP or conditioned place aversion in male rats (n = 6) or in female rats (n = 6). CONCLUSIONS: The results of this study show that Ro 63-1908 is able to decrease the conditioned rewarding effects of methamphetamine.

The association between risky decision making and cocaine conditioned place preference is moderated by sex.

BACKGROUND: Excessive risk taking is a characteristic trait of several psychiatric conditions, including substance use disorders. High risk-taking (HiR) rats self-administer more cocaine compared to low risk-taking (LoR) rats. However, research has not determined if risk taking is associated with enhanced cocaine conditioned place preference (CPP). METHODS: Male and female Sprague Dawley rats (n = 48 each sex) were first tested in the risky decision task (RDT), in which a response on one lever resulted in safe delivery of one food pellet, and a response on a different lever resulted in delivery of two pellets and probabilistic delivery of foot shock. Following RDT training, rats were tested for cocaine CPP. The first session was a pretest that measured rats' preference for three compartments that provided different visual and tactile cues. Rats then learned to associate one compartment with cocaine (either 10.0 mg/kg or 20.0 mg/kg; i.p.) and one compartment with saline (1.0 ml/kg; i.p.) across eight conditioning sessions. Finally, rats explored all three compartments in a drug-free state. RESULTS: Sex significantly moderated the association between risky decision making and cocaine CPP. While increased risk aversion was somewhat positively associated with cocaine CPP in males, increased risk taking was positively correlated with cocaine CPP in females. CONCLUSIONS: These results highlight the moderating role of sex on the relationship between risky decision making and cocaine reward.

Differential effects of glutamate N-methyl-D-aspartate receptor antagonists on risky choice as assessed in the risky decision task.

RATIONALE: Risky choice can be measured using the risky decision task (RDT). In the RDT, animals choose between a large, risky option that is paired with probabilistic foot shock and a small, safe option that is never paired with shock. To date, studies examining the neurochemical basis of decision-making in the RDT have focused primarily on the dopaminergic system but have not focused on the glutamatergic system, which has been implicated in risky decision-making. OBJECTIVES: Because glutamate is known to play a critical role in decision-making, we wanted to determine the contribution of the glutamatergic system to performance in the RDT. METHODS: In the experiment, 32 rats (16 male; 16 female) were tested in the RDT. The probability of receiving a foot shock increased across the session (ascending schedule) for half of the rats but decreased across the session (descending schedule) for half of the rats. Following training, rats received injections of the N-methyl-D-aspartate (NMDA) receptor competitive antagonist CGS 19755 (0, 1.0, 2.5, 5.0 mg/kg; s.c.) and the GluN2B-selective antagonist Ro 63-1908 (0, 0.1, 0.3, 1.0 mg/kg; s.c.). RESULTS: CGS 19755 (2.5 and 5.0 mg/kg) increased risky choice in males and females trained on the ascending schedule. Ro 63-1908 (1.0 mg/kg) decreased risky choice, but only in male rats trained on the ascending schedule. CONCLUSIONS: Although NMDA receptor antagonists differentially alter risky choice in the RDT, the current results show that NMDA receptors are an important mediator of decision-making involving probabilistic delivery of positive punishment.