Large Language Models in Medical Education: Comparing ChatGPT- to Human-Generated Exam Questions.

PROBLEM: Creating medical exam questions is time consuming, but well-written questions can be used for test-enhanced learning, which has been shown to have a positive effect on student learning. The automated generation of high-quality questions using large language models (LLMs), such as ChatGPT, would therefore be desirable. However, there are no current studies that compare students' performance on LLM-generated questions to questions developed by humans. APPROACH: The authors compared student performance on questions generated by ChatGPT (LLM questions) with questions created by medical educators (human questions). Two sets of 25 multiple-choice questions (MCQs) were created, each with 5 answer options, 1 of which was correct. The first set of questions was written by an experienced medical educator, and the second set was created by ChatGPT 3.5 after the authors identified learning objectives and extracted some specifications from the human questions. Students answered all questions in random order in a formative paper-and-pencil test that was offered leading up to the final summative neurophysiology exam (summer 2023). For each question, students also indicated whether they thought it had been written by a human or ChatGPT. OUTCOMES: The final data set consisted of 161 participants and 46 MCQs (25 human and 21 LLM questions). There was no statistically significant difference in item difficulty between the 2 question sets, but discriminatory power was statistically significantly higher in human than LLM questions (mean = .36, standard deviation [SD] = .09 vs mean = .24, SD = .14; P = .001). On average, students identified 57% of question sources (human or LLM) correctly. NEXT STEPS: Future research should replicate the study procedure in other contexts (e.g., other medical subjects, semesters, countries, and languages). In addition, the question of whether LLMs are suitable for generating different question types, such as key feature questions, should be investigated.

Treatment of peripheral pain with low-dose local anesthetics by epidermal, epithelial and periosteal application.

OBJECTIVES: The efficiency of local anesthetics (LAs) in the treatment of peripheral pain is commonly attributed to their capacity to block the axon conduction of sensory nerves. LAs even in non-blocking concentration suppress oscillations of the resting membrane potential. Spiking in sensory neurons is triggered by subthreshold membrane potential oscillations (SMPOs), which reach threshold and is maintained by depolarizing impulse after oscillations. The suppression of these oscillations abolishes sustained afferent discharge in sensory nerves without blocking the axon conduction. In a retrospective observational study, we examined if LAs in low concentration and very small quantities could reduce peripheral pain in patients. DESIGN: During a period of 2 years, a total of 127 consecutive patients, 43 with cervico-brachial, 12 with intercostal and 72 with lumbo-sciatic pain received an identical treatment, which consisted of LAs applied in 4-8 sessions on average to a fixed set of epidermal, epithelial and periosteal locations. The primary outcome was relief of symptoms measured by verbal analog scales at the end of therapy. RESULTS: At the end of therapy, 53 (41.7%) of all patients (127) had a complete remission (reduction of pain 100%). Twenty-three patients (18.1%) had a partial remission with >90% reduction of pain and 50 patients (39.4%) had a pain reduction of 30%-90%. One patient did not respond. CONCLUSION: LAs in low concentration and small quantities proved to be highly efficient in the treatment of peripheral pain. An almost complete remission could be obtained in a majority of patients. Given the extent of pain reduction achieved, the method of application seems to be of major importance.

Modulation of the Ca2+ conductance of nicotinic acetylcholine receptors by Lypd6.

The agonist binding sensitivity and desensitization kinetics of nicotinic acetylcholine receptors (nAChRs) can be modulated by snake venom neurotoxins and related endogenous small proteins of the uPAR-Ly6 family. Here we identify Lypd6, a distantly related member of the u-PAR/Ly-6 family expressed in neurons as a novel modulator of nAChRs. Lypd6 overexpressed in trigeminal ganglia neurons selectively enhanced the Ca2+-component of nicotine-evoked currents through nAChRs, as evidenced by comparative whole-cell patch clamp recordings and Ca2+-imaging in wildtype and transgenic mice overexpressing Lypd6. In contrast, a knockdown of Lypd6 expression using siRNAs selectively reduced nicotine-evoked Ca2+-currents. Pharmacological experiments revealed that the nAChRs involved in this process are heteromers. Transgenic mice displayed behaviors that were indicative of an enhanced cholinergic tone, such as a higher locomotor arousal, increased prepulse-inhibition and hypoalgesia. These mice overexpressing Lypd6 mice were also more sensitive to the analgesic effects of nicotine. Transgenic mice expressing siRNAs directed against Lypd6 were unable to procreate, thus indicating a vital role for this protein. Taken together, Lypd6 seems to constitute a novel modulator of nAChRs that affects receptor function by selectively increasing Ca2+-influx through this ion channels.

Reversal of pathological pain through specific spinal GABAA receptor subtypes.

Inflammatory diseases and neuropathic insults are frequently accompanied by severe and debilitating pain, which can become chronic and often unresponsive to conventional analgesic treatment. A loss of synaptic inhibition in the spinal dorsal horn is considered to contribute significantly to this pain pathology. Facilitation of spinal gamma-aminobutyric acid (GABA)ergic neurotransmission through modulation of GABA(A) receptors should be able to compensate for this loss. With the use of GABA(A)-receptor point-mutated knock-in mice in which specific GABA(A) receptor subtypes have been selectively rendered insensitive to benzodiazepine-site ligands, we show here that pronounced analgesia can be achieved by specifically targeting spinal GABA(A) receptors containing the alpha2 and/or alpha3 subunits. We show that their selective activation by the non-sedative ('alpha1-sparing') benzodiazepine-site ligand L-838,417 (ref. 13) is highly effective against inflammatory and neuropathic pain yet devoid of unwanted sedation, motor impairment and tolerance development. L-838,417 not only diminished the nociceptive input to the brain but also reduced the activity of brain areas related to the associative-emotional components of pain, as shown by functional magnetic resonance imaging in rats. These results provide a rational basis for the development of subtype-selective GABAergic drugs for the treatment of chronic pain, which is often refractory to classical analgesics.

Spinal inflammatory hyperalgesia is mediated by prostaglandin E receptors of the EP2 subtype.

Blockade of prostaglandin (PG) production by COX inhibitors is the treatment of choice for inflammatory pain but is also prone to severe side effects. Identification of signaling elements downstream of COX inhibition, particularly of PG receptor subtypes responsible for pain sensitization (hyperalgesia), provides a strategy for better-tolerated analgesics. Here, we have identified PGE2 receptors of the EP2 receptor subtype as key signaling elements in spinal inflammatory hyperalgesia. Mice deficient in EP2 receptors (EP2-/- mice) completely lack spinal PGE2-evoked hyperalgesia. After a peripheral inflammatory stimulus, EP2-/- mice exhibit only short-lasting peripheral hyperalgesia but lack a second sustained hyperalgesic phase of spinal origin. Electrophysiological recordings identify diminished synaptic inhibition of excitatory dorsal horn neurons as the dominant source of EP2 receptor-dependent hyperalgesia. Our results thus demonstrate that inflammatory hyperalgesia can be treated by targeting of a single PG receptor subtype and provide a rational basis for new analgesic strategies going beyond COX inhibition.

Glycinergic neurons expressing enhanced green fluorescent protein in bacterial artificial chromosome transgenic mice.

Although glycine is a major inhibitory transmitter in the mammalian CNS, the role of glycinergic neurons in defined neuronal circuits remains ill defined. This is due in part to difficulties in identifying these cells in living slice preparations for electrophysiological recordings and visualizing their axonal projections. To facilitate the morphological and functional analysis of glycinergic neurons, we generated bacterial artificial chromosome (BAC) transgenic mice, which specifically express enhanced green fluorescent protein (EGFP) under the control of the promotor of the glycine transporter (GlyT) 2 gene, which is a reliable marker for glycinergic neurons. Neurons expressing GlyT2-EGFP were intensely fluorescent, and their dendrites and axons could be visualized in great detail. Numerous positive neurons were detected in the spinal cord, brainstem, and cerebellum. The hypothalamus, intralaminar nuclei of the thalamus, and basal forebrain also received a dense GlyT2-EGFP innervation, whereas in the olfactory bulb, striatum, neocortex, hippocampus, and amygdala positive fibers were much less abundant. No GlyT2-EGFP-positive cell bodies were seen in the forebrain. On the subcellular level, GlyT2-EGFP fluorescence was colocalized extensively with glycine immunoreactivity in somata and dendrites and with both glycine and GlyT2 immunoreactivity in axon terminals, as shown by triple staining at all levels of the neuraxis, confirming the selective expression of the transgene in glycinergic neurons. In slice preparations of the spinal cord, no difference between the functional properties of EGFP-positive and negative neurons could be detected, confirming the utility of visually identifying glycinergic neurons to investigate their functional role in electrophysiological studies.

The AMPA receptor subunits GluR-A and GluR-B reciprocally modulate spinal synaptic plasticity and inflammatory pain.

Ca(2+)-permeable AMPA receptors are densely expressed in the spinal dorsal horn, but their functional significance in pain processing is not understood. By disrupting the genes encoding GluR-A or GluR-B, we generated mice exhibiting increased or decreased numbers of Ca(2+)-permeable AMPA receptors, respectively. Here, we demonstrate that AMPA receptors are critical determinants of nociceptive plasticity and inflammatory pain. A reduction in the number of Ca(2+)-permeable AMPA receptors and density of AMPA channel currents in spinal neurons of GluR-A-deficient mice is accompanied by a loss of nociceptive plasticity in vitro and a reduction in acute inflammatory hyperalgesia in vivo. In contrast, an increase in spinal Ca(2+)-permeable AMPA receptors in GluR-B-deficient mice facilitated nociceptive plasticity and enhanced long-lasting inflammatory hyperalgesia. Thus, AMPA receptors are not mere determinants of fast synaptic transmission underlying basal pain sensitivity as previously thought, but are critically involved in activity-dependent changes in synaptic processing of nociceptive inputs.

GlyR alpha3: an essential target for spinal PGE2-mediated inflammatory pain sensitization.

Prostaglandin E2 (PGE2) is a crucial mediator of inflammatory pain sensitization. Here, we demonstrate that inhibition of a specific glycine receptor subtype (GlyR alpha3) by PGE2-induced receptor phosphorylation underlies central inflammatory pain sensitization. We show that GlyR alpha3 is distinctly expressed in superficial layers of the spinal cord dorsal horn. Mice deficient in GlyR alpha3 not only lack the inhibition of glycinergic neurotransmission by PGE2 seen in wild-type mice but also show a reduction in pain sensitization induced by spinal PGE2 injection or peripheral inflammation. Thus, GlyR alpha3 may provide a previously unrecognized molecular target in pain therapy.

Gabapentin may inhibit synaptic transmission in the mouse spinal cord dorsal horn through a preferential block of P/Q-type Ca2+ channels.

Gabapentin is a lipophilic analog of gamma-amino butyric acid (GABA) with therapeutic activity against certain forms of epilepsy and neuropathic pain. Despite its structural similarity to GABA, it does not bind GABAA or GABAB receptors and the mechanism, especially of its analgesic action, has remained elusive. Here, we have studied its effects on synaptic transmission mediated by the major spinal fast excitatory and inhibitory neurotransmitters, L-glutamate and glycine, in the superficial layers of the spinal cord dorsal horn, a CNS area, which is critically involved in nociception. Gabapentin reversibly reduced evoked excitatory postsynaptic currents mediated by alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA-EPSCs) and inhibitory postsynaptic currents mediated by glycine (gly-IPSCs). Inhibition of AMPA-EPSCs and gly-IPSCs occurred with similar potencies (approximately 10-50 nM) and by about the same degree (approximately 40% at 1 microM). Gabapentin did not affect membrane currents elicited by exogenously applied glutamate or glycine arguing against a postsynaptic site of action. Selective blockade of N-type Ca2+ channels with omega-conotoxin GVIA dramatically increased and blockade of P/Q-type channels with omega-agatoxin IVA strongly attenuated inhibition of evoked synaptic transmission by gabapentin. These results show that gabapentin affects both excitatory and inhibitory spinal neurotransmission via a presynaptic mechanism which preferentially involves P/Q-type Ca2+ channels.

Facilitation of spinal NMDA receptor currents by spillover of synaptically released glycine.

In the mammalian CNS, N-methyl-D-aspartate (NMDA) receptors serve prominent roles in many physiological and pathophysiological processes including pain transmission. For full activation, NMDA receptors require the binding of glycine. It is not known whether the brain uses changes in extracellular glycine to modulate synaptic NMDA responses. Here, we show that synaptically released glycine facilitates NMDA receptor currents in the superficial dorsal horn, an area critically involved in pain processing. During high presynaptic activity, glycine released from inhibitory interneurons escapes the synaptic cleft and reaches nearby NMDA receptors by so-called spillover. In vivo, this process may contribute to the development of inflammatory hyperalgesia.

PGE(2) selectively blocks inhibitory glycinergic neurotransmission onto rat superficial dorsal horn neurons.

Despite the crucial role that prostaglandins (PGs) have in the sensitization of the central nervous system to pain, their cellular and molecular targets leading to increased pain perception have remained elusive. Here we investigated the effects of PGE(2) on fast synaptic transmission onto neurons in the rat spinal cord dorsal horn, the first site of synaptic integration in the pain pathway. We identified the inhibitory (strychnine-sensitive) glycine receptor as a specific target of PGE(2). PGE(2), but not PGF(2 alpha), PGD(2) or PGI(2), reduced inhibitory glycinergic synaptic transmission in low nanomolar concentrations, whereas GABAA, AMPA and NMDA receptor-mediated transmission remained unaffected. Inhibition of glycine receptors occurred via a postsynaptic mechanism involving the activation of EP2 receptors, cholera-toxin-sensitive G-proteins and cAMP-dependent protein kinase. Via this mechanism, PGE(2) may facilitate the transmission of nociceptive input through the spinal cord dorsal horn to higher brain areas where pain becomes conscious.