Ramakrishnan Arjun, Hayden Benjamin Y, Platt Michael L
Department of Neuroscience, University of Pennsylvania, Philadelphia, PA, USA.
Department of Neuroscience, University of Minnesota, Minneapolis, MN, USA.
Brain Neurosci Adv. 2019 Jan 18;3:2398212818817932. doi: 10.1177/2398212818817932. eCollection 2019 Jan-Dec.
To maximise long-term reward rates, foragers deciding when to leave a patch must compute a decision variable that reflects both the immediately available reward and the time costs associated with travelling to the next patch. Identifying the mechanisms that mediate this computation is central to understanding how brains implement foraging decisions. We previously showed that firing rates of dorsal anterior cingulate sulcus neurons incorporate both variables. This result does not provide information about whether integration of information reflected in dorsal anterior cingulate sulcus spiking activity arises locally or whether it is inherited from upstream structures. Here, we examined local field potentials gathered simultaneously with our earlier recordings. In the majority of recording sites, local field potential spectral bands - specifically theta, beta, and gamma frequency ranges - encoded immediately available rewards but not time costs. The disjunction between information contained in spiking and local field potentials can constrain models of foraging-related processing. In particular, given the proposed link between local field potentials and inputs to a brain area, it raises the possibility that local processing within dorsal anterior cingulate sulcus serves to more fully bind immediate reward and time costs into a single decision variable.
为了使长期奖励率最大化,决定何时离开一个斑块的觅食者必须计算一个决策变量,该变量既要反映即时可得的奖励,也要反映前往下一个斑块所涉及的时间成本。确定介导这种计算的机制对于理解大脑如何做出觅食决策至关重要。我们之前表明,背侧前扣带回沟神经元的放电率包含了这两个变量。这一结果并未提供有关背侧前扣带回沟尖峰活动所反映的信息整合是在局部发生,还是从上游结构继承而来的信息。在这里,我们检查了与我们早期记录同时收集的局部场电位。在大多数记录位点,局部场电位频谱带——特别是theta、beta和gamma频率范围——编码了即时可得的奖励,但没有编码时间成本。尖峰电位和局部场电位所包含信息之间的脱节可以限制觅食相关处理的模型。特别是,考虑到局部场电位与大脑区域输入之间的假定联系,这增加了一种可能性,即背侧前扣带回沟内的局部处理有助于将即时奖励和时间成本更充分地结合成一个单一的决策变量。