Zhang Zhi, Bodznick David
Biology Department, Wesleyan University, Middletown, CT 06459, USA.
J Exp Biol. 2008 Dec;211(Pt 23):3720-8. doi: 10.1242/jeb.020099.
Detection of relevant sensory signals requires the filtering out of irrelevant noise, including noise created by the animal's own movements (reafference). This is accomplished in the electrosense of little skates (Raja erinacea) by an adaptive filter in the cerebellar-like electrosensory nucleus (dorsal nucleus) in the medulla. We have shown that electrosensory inputs reliably coupled to the regularly recurring movements of breathing over time are eliminated selectively in the principal neurons (ascending efferent neurons, AENs) by a cancellation signal that is a negative of the reafference and is supplied by a parallel fiber system. Similarly, electrosensory inputs repeatedly linked to passive fin movements are eliminated suggesting that the filter also functions in relation to other behaviors besides breathing. To determine whether this adaptive filter can eliminate reafference created by brief and infrequent episodic behaviors like swimming in skates, we initiated a series of coupling tests in which an external electrosensory stimulus was coupled to short bouts of either parallel fiber stimulation or passive fin movements, and then measured the ability of AENs to generate a cancellation signal. Following five brief coupling periods (30-60 s) separated by long rest periods (1-9 min), 38.5% of the AENs developed a cancellation signal when the coupling was to parallel fiber stimulation, and 73% when the coupling was to passive fin movement. We demonstrate that the cancellation signals can be developed incrementally, persist for at least a 3 h rest period without reinforcement, and are extinguished within minutes when the association of sensory stimulus and fin movement or parallel fiber stimulation no longer exists. The results indicate that the adaptive filter has the properties necessary to cancel reafference associated with even brief and infrequent behaviors.
检测相关的感觉信号需要滤除无关的噪声,包括动物自身运动产生的噪声(再传入)。在小鳐(Raja erinacea)的电感觉系统中,这是通过延髓中类似小脑的电感觉核(背核)中的自适应滤波器来实现的。我们已经表明,随着时间的推移,可靠地与有规律重复的呼吸运动耦合的电感觉输入,在主要神经元(升支传出神经元,AENs)中被一个抵消信号选择性地消除,该抵消信号是再传入信号的负值,由平行纤维系统提供。同样,反复与被动鳍运动相关联的电感觉输入也被消除,这表明该滤波器在除呼吸之外的其他行为中也起作用。为了确定这种自适应滤波器是否能够消除由小鳐短暂且不频繁的偶发行为(如游泳)产生的再传入信号,我们启动了一系列耦合测试,其中将外部电感觉刺激与短时间的平行纤维刺激或被动鳍运动耦合,然后测量AENs产生抵消信号的能力。在由长时间休息期(1 - 9分钟)隔开的五个短暂耦合期(30 - 60秒)之后,当耦合为平行纤维刺激时,38.5%的AENs产生了抵消信号,当耦合为被动鳍运动时,这一比例为73%。我们证明,抵消信号可以逐渐形成,在没有强化的情况下至少持续3小时的休息期,并且当感觉刺激与鳍运动或平行纤维刺激之间的关联不再存在时,几分钟内就会消失。结果表明,自适应滤波器具有消除与即使是短暂且不频繁行为相关的再传入信号所需的特性。