Ensemble accumulator modeling of speed-accuracy tradeoff in visual search

 描述(由申请人提供)：本提案的目标是开展研究性培训，以开发受神经生理学启发和约束的认知建模方面的新专业知识。我之前在人类群体决策方面的工作，以及发起人最近提出的探索神经集成决策的新的建模框架，为新的培训和跨学科合作提供了一个极好的机会。目的1系统地研究决策选择任务中被实例化为证据累积器的神经元和神经累积器集合的集体行为之间的关系；预测的选择行为和集合的动力学将与观察到的眼跳选择行为和清醒行为猴子的测量神经动力学进行比较。目标2将通过探索集合中累加器之间聚合的替代计算机制，以及从群决策的其他领域中提取替代聚合方案的假设，来显著扩展这项工作。培训潜力很大，因为这项工作将为神经元集合做出的决策开发新的计算模型，这是一个我以前没有经验的研究领域，但它提供了令人兴奋的新方向，可以链接到我之前在人群智慧方面的工作。在我新的博士后工作和之前的博士工作之间建立新的联系，将为我未来作为一名独立研究员的研究奠定基础。这些发现具有潜在的广泛兴趣和适用性，因为它可能有助于深入了解神经元集合是如何在大脑中共同做出决定的。

 DESCRIPTION (provided by applicant): The goal of this proposal is to carry out research training to develop new expertise in cognitive modeling that is inspired and constrained by neurophysiology. My previous work in human group decision making, and the sponsors' recent formulation of a novel modeling framework to explore neural ensemble decision making, yields an excellent opportunity for new training and interdisciplinary collaboration. Aim 1 will systematically investigate the relationship between neurons, instantiated as evidence accumulators, and the collective behavior of ensembles of neural accumulators in decision choice tasks; predicted choice behavior and dynamics of ensembles will be compared to observed saccade choice behavior and measured neural dynamics of awake behaving monkeys. Aim 2 will significantly extend this work by exploring alternative computational mechanisms for aggregation across accumulators in an ensemble, with hypotheses for alternative aggregation schemes drawn from other domains of group decision making. Training potential is high because the work will develop new computational models of decisions made by ensembles of neurons, an area of research in which I have no prior experience, but which offers exciting new directions for linking to my prior work on the Wisdom of Crowds. Forging novel connections between my new postdoctoral work and my previous PhD work will form a basis for my future research as an independent investigator. The findings have potential wide interest and applicability, as it may shed insight into how ensembles of neurons collectively make decisions in the brain.