Research report
Baseline hippocampal theta oscillation speeds correlate with rate of operant task acquisition

https://doi.org/10.1016/j.bbr.2008.02.032Get rights and content

Abstract

Many lines of evidence indicate that theta rhythm, a prominent neural oscillatory mode found in the mammalian hippocampus, plays a key role in the acquisition, processing, and retrieval of memories. However, a predictive neurophysiological feature of the baseline theta rhythm that correlates with the learning rate across different animals has yet to be identified. Here we show that the mean theta rhythm speed observed during baseline periods of immobility has a strong positive correlation with the rate at which rats learn an operant task. This relationship is observed across rats, during both quiet waking (r = 0.82; p < 0.01) and paradoxical sleep (r = 0.83; p < 0.01), suggesting that the basal theta frequency relates to basic neurological processes that are important in the acquisition of operant behavior.

Introduction

Hippocampal theta oscillations (HTOs) are strongly synchronized electroencephalograms generated by the brain, and occur in the 4–11 Hz frequency range [1], [2]. Previous research has provided clear evidence that HTOs are triggered by novel stimulation [3], and several types of learning such as operant conditioning [4], [5]. While HTOs during waking (WK) are typically accompanied by increased alertness [1], [6], [7], they are also a typical marker of paradoxical sleep (PS; also known as rapid eye movement sleep). Several lines of evidence suggest that HTOs play a prominent role in the acquisition and processing of new memories. Theta oscillations are always present across hippocampus when animals explore a novel environment [3], [8]. Additionally, neuronal populations across hippocampus and prefrontal cortex involved in spatial memory formation, and local field potentials across amygdala involved in fear conditioning and the formation of emotional memories, phase-lock to HTOs [9], [10]. Furthermore, septal lesions which abolish HTOs are associated with profound anterograde amnesia [11]. Together these findings implicate this form of neural oscillation as an important modulator of learning processes.

Despite the compelling evidence supporting HTOs as a critical modulator of information storage, the exact nature of the relationship between this oscillatory neural mode and behavioral learning rates has yet to be determined. Several studies using concurrent electrophysiological recordings have shown that HTO power increases with task difficulty [12] and decreases as learning progresses [13]. Moreover, one classical study utilizing electrophysiological data recorded immediately prior to task performance showed that HTO power predicts learning rates across animals [14]. Unfortunately, these studies did not temporally dissociate their electrophysiological recordings from periods of behavioral task performance. Given that both HTO power and learning rates vary with attention state [15], [16], additional study is necessary in order to elucidate the relationship between basal HTO properties and learning rates. Here we demonstrate that the mean frequency of HTOs during baseline periods positively correlates with learning rates across rodents performing a long-term operant conditioning learning task. Moreover, this relationship is observed during both quiet waking and PS. Thus, our findings provide the first experimental evidence correlating the mean frequency of hippocampal theta oscillations with learning rates across different animals.

Section snippets

Animal care and use

All experimental procedures were designed to avoid discomfort of the animals, and were in strict accordance with Brazilian Federal Animal Welfare of Health Guidelines and the University of Sao Paulo Animal Care and Use Committee on animal well being. Male Whistar rats (n = 9) weighing a mean of 380 g were used for all experiments presented. Rats were housed in individual cages at 20–25 °C, in a 12/12 h light/dark cycle. Food restriction was initiated 5 days before training; animals were fed 20 g of

Results

Nine male Whistar rats were subjected to an operant behavioral learning task across 15 days. As expected, learning occurred in a stepwise manner over several daily 1-h sessions, with considerable variation among animals. For example, the fastest learner reached >90% of correct performance in as little as 9 days, while the slowest learner failed to reach the same criterion even after 15 days of continuous training (Fig. 1). The slope of the learning curve was determined by linear regression and

Discussion

Theta rhythm patterns were first recorded in rabbits in 1938 by Jung and Kornmüller [20], and have since been identified and studied extensively in many different species, including humans [5], [12], [21], [22], [23]. Nevertheless, the neurophysiological role of theta oscillations remains largely unknown. One classic study successfully demonstrated that learning rates correlated with normalized HTO power observed prior to task performance [14]; however, this study did not temporally dissociate

Acknowledgments

We thank Dr. György Buzsáki for insightful comments and critical reading of this manuscript. Work supported by NIH and FAPESP grant.

References (29)

  • C. Timo-Iaria et al.

    Physiol Behav

    (1970)
  • G. Buzsaki

    Neuron

    (2002)
  • W.R. Adey

    Prog Brain Res

    (1967)
  • E. Grastyan et al.

    Electroencephalogr Clin Neurophysiol Suppl

    (1959)
  • F.H. Lopes da Silva et al.

    Electroencephalogr Clin Neurophysiol

    (1969)
  • C.H. Vanderwolf

    Electroencephalogr Clin Neurophysiol

    (1969)
  • J. Winson

    Behav Biol

    (1972)
  • A.G. Siapas et al.

    Neuron

    (2005)
  • T. Masuoka et al.

    Brain Res

    (2006)
  • S. Grossberg

    Conscious Cogn

    (1999)
  • G. Buzsaki et al.

    Brain Res

    (1983)
  • P. Fries et al.

    Trends Neurosci

    (2007)
  • J.D. Green et al.

    J Neurophysiol

    (1954)
  • T. Seidenbecher et al.

    Science

    (2003)
  • Cited by (10)

    • The relationship between sleep and appetitive conditioning: A systematic review and meta-analysis

      2023, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      In five of the thirteen studies, REM sleep duration and number of REM episodes were significantly increased following conditioning in rats (Lecas, 1976; Leenaars et al., 2013; Smith et al., 1977; Tripathi and Jha, 2016; Tripathi et al., 2018), with these effects lasting up to three (Leenaars et al., 2013) and four hours post-conditioning (Tripathi and Jha, 2016; Tripathi et al., 2018). Similarly, other findings showed performance on tasks was correlated with: a) amount of REM sleep (Tripathi et al., 2018), and b) mean hippocampal theta oscillations during REM sleep (Santos et al., 2008). There is also evidence suggesting neural patterns important for learning operant behaviours during an active period re-emerge during REM sleep (i.e., neural replay; Marczynski et al., 1992) and that task performance is associated with similar temporal characteristics as REM sleep (e.g., proportion of time spent REM sleep vs. performance; Sterman et al., 1972).

    • Instrumental learning: An animal model for sleep dependent memory enhancement

      2013, Journal of Neuroscience Methods
      Citation Excerpt :

      In contrast to some previous studies, and in contrast to our first hypothesis, we did not find any evidence for baseline differences in sleep parameters between fast-learning and slow-learning rats. Only one rodent study describes an association of instrumental learning with hippocampal theta oscillation speed (Santos et al., 2008). These findings may be inconclusive because hippocampal activity was determined after task acquisition rather than at true baseline, and may thus have been influenced by the learning experience.

    View all citing articles on Scopus
    1

    Deaseased (1924–2005).

    View full text