Deadwyler S A, Bunn T, Hampson R E
Department of Physiology and Pharmacology, Bowman Gray School of Medicine, Wake Forest University, Winston-Salem, North Carolina 27157, USA.
J Neurosci. 1996 Jan;16(1):354-72. doi: 10.1523/JNEUROSCI.16-01-00354.1996.
Multiple-cell recording from specially designed arrays of microwire electrodes allowed analysis of anatomically defined ensemble activity from 10 different locations within the hippocampus of rats (n = 7) performing a two-lever operant version of a spatial delayed-nonmatch-to-sample task (DNMS). Application of population analysis procedures to ensembles of single-neuron activity within the CA1 and CA3 fields revealed firing patterns related to task-relevant events within a DNMS trial. The patterns were extracted via a canonical discriminant analysis in the form of "roots" that represented sources of variance in firing within the ensemble, such as phase of the task (Sample or Nonmatch), spatial position of the lever press response (left or right), and correct versus error trials. Comparison of the ensemble firing on correct versus error trials revealed important insight into ensemble information encoding, such as "miscoding" of the response position and lack of distinct encoding of the response in the Sample phase, which became increasingly vulnerable to error as a function of the duration of delay interval. The extracted discriminant scores were reflective of multiple representations within ensembles and suggested that "conjunctions" of task-relevant features could be represented effectively by small numbers of hippocampal neurons. The findings support the long-held supposition that hippocampal neurons play a critical role in the encoding and retrieval of information in recognition memory tasks.
从专门设计的微丝电极阵列进行的多细胞记录,使得对执行空间延迟非匹配样本任务(DNMS)的两杆操作性版本的大鼠(n = 7)海马体中10个不同位置的解剖学定义的整体活动进行分析成为可能。将群体分析程序应用于CA1和CA3区域内的单个神经元活动群体,揭示了与DNMS试验中与任务相关事件相关的放电模式。这些模式通过典型判别分析以“根”的形式提取,这些“根”代表了群体内放电变化的来源,例如任务阶段(样本或非匹配)、杠杆按压反应的空间位置(左或右)以及正确与错误试验。正确试验与错误试验的整体放电比较揭示了对整体信息编码的重要见解,例如反应位置的“错误编码”以及样本阶段反应缺乏明显编码,并随着延迟间隔持续时间的增加而越来越容易出错。提取的判别分数反映了群体内的多种表征,并表明与任务相关特征的“结合”可以由少量海马神经元有效表征。这些发现支持了长期以来的假设,即海马神经元在识别记忆任务中的信息编码和检索中起着关键作用。