Theta-Contingent Trial Presentation Accelerates Learning Rate and Enhances Hippocampal Plasticity During Trace Eyeblink Conditioning

Behavioral Neuroscience Copyright 2004 by the American Psychological Association 2004, Vol. 118, No. 2, 403 411 0735-7044/04/$12.00 DOI: 10.1037/0735-7044.118.2.403 Theta-Contingent Trial Presentation Accelerates Learning Rate and Enhances Hippocampal Plasticity During Trace Eyeblink Conditioning Amy L. Griffin, Yukiko Asaka, Ryan D. Darling, and Stephen D. Berry Miami University Hippocampal theta activity has been established as a key predictor of acquisition rate in rabbit (Orcytolagus cuniculus) classical conditioning. The current study used an online brain computer interface to administer conditioning trials only in the explicit presence or absence of spontaneous theta activity in the hippocampus-dependent task of trace conditioning. The findings indicate that animals given theta-contingent training learned significantly faster than those given nontheta-contingent training. In parallel with the behavioral results, the theta-triggered group, and not the nontheta-triggered group, exhibited profound increases in hippocampal conditioned unit responses early in training. The results not only suggest that theta-contingent training has a dramatic facilitory effect on trace conditioning but also implicate theta activity in enhancing the plasticity of hippocampal neurons. There is little doubt that the hippocampus plays a key role in learning and memory processes. However, the precise nature of this involvement is the subject of much debate (for reviews, see Eichenbaum, Otto, & Cohen, 1992; Redish, 2001). One avenue of research has focused on the relationship between cognition and one of the most striking and well-studied indices of hippocampal activity, the theta rhythm. The theta rhythm is a 3 12-Hz sinusoidal-like waveform found in the hippocampus and related structures during a variety of cognitive behaviors in several species (Adey, 1966; Asaka, Griffin, & Berry, 2002; Asaka, Seager, Griffin, & Berry, 2000; Bennett, 1969; Berry, Seager, Asaka, & Borgnis, 2000; Berry & Thompson, 1978; Givens & Olton, 1990; Green & Arduini, 1954; Kaneko & Thompson, 1997; Lawson & Bland, 1993; Markowska, Olton, & Givens, 1995; Mizumori, Perez, Alvarado, Barnes, & McNaughton, 1990; Salvatierra & Berry, 1989; Seager, Johnson, Chabot, Asaka, & Berry, 2002; Tracy, Jarrard, & Davidson, 2001; Vanderwolf, 1971; Winson, 1978), including humans (Arnolds, Lopes da Silva, Aitink, Kamp, & Boeijinga, 1980; Kahana, Seeling, & Madsen, 2001; Meador et al., 1991; Raghavachari et al., 2001; Tesche & Karhu, 2000). Although many of these investigations have demonstrated a strong relationship between the presence of hippocampal theta activity Amy L. Griffin, Yukiko Asaka, Ryan D. Darling, and Stephen D. Berry, Department of Psychology and Center for Neuroscience, Miami University. Amy L. Griffin is now at the Center for Memory and Brain, Boston University. Yukiko Asaka is now at the Department of Psychiatry, Yale University School of Medicine. This project was supported by a Sigma Xi grant to Yukiko Asaka. We would like to thank Elizabeth Shurell, Bradley Haverkos, Rachel Vesco, Kevin Moran, and Cristina Hochwalt for their assistance with data collection and analysis. We would also like to thank Lynn Johnson for his technical assistance and Matthew Seager for his helpful suggestions on an earlier version of this article. Correspondence concerning this article should be addressed to Stephen D. Berry, Department of Psychology, Miami University, 216 Benton Hall, Oxford, OH 45056. E-mail: berrysd@muohio.edu and learning rate, most have used lesion and pharmacological techniques to permanently alter the activity of the hippocampus or of its afferents in order to disrupt or enhance learning. Our laboratory has recently developed a less invasive approach to demonstrate the relationship between theta and learning: monitoring naturally occurring theta activity and administering training trials contingent on its presence or absence. This not only ensures normal, physiological patterns of theta activity, but also permits the typical, but potentially significant, alternations between theta and nontheta that occur in undrugged intact animals. In a recent study using a delay eyeblink conditioning paradigm, Seager et al. (2002) found that rabbits that received training trials in the presence of theta learned twice as fast as those receiving trials in the absence of theta. To control for differences in intertrial interval (ITI) length and number of trials per session, each theta- and nontheta-triggered animal was assigned a yoked control animal that received the same ITIs and trials per day as its partner, irrespective of the amount of theta activity present before each trial. The results showed that theta-triggered animals learned marginally faster than their yoked control group, but that nonthetatriggered animals learned significantly slower than theta-triggered and both yoked control groups. This finding suggests that instead of theta benefiting learning, nontheta is especially detrimental, at least in tasks that do not require the hippocampus, such as delay conditioning. These findings corroborated and extended work begun in the 1970s that demonstrated a strong predictive relationship between the amount of theta activity and subsequent learning rate in delay nictitating membrane conditioning (Berry & Swain, 1989; Berry & Thompson, 1978, 1979). However, questions remain about the relative impact of theta versus nontheta on hippocampusdependent tasks and the relation between theta states and conditioned unit responses in hippocampus. The objective of the current study was to evaluate theta-related modulation of both the early and asymptotic stages of acquisition in trace eyeblink conditioning, a paradigm in which there is no overlap between the conditioned stimulus (CS) and unconditioned stimulus (US). We expected to observe behavioral differences between the theta- and nontheta-triggered groups similar to those 403

404 GRIFFIN, ASAKA, DARLING, AND BERRY seen in the Seager et al. (2002) study, with the theta-triggered group learning significantly faster than the nontheta-triggered group. Whereas the cerebellum and related pathways are essential to both delay and trace forms of eyeblink conditioning, hippocampal influences seem to be parallel and modulatory, becoming essential in more complex paradigms such as the trace paradigm. Therefore, unlike the Seager et al. (2002) delay conditioning study, we expected to see differences in the earliest stage of learning (defined here as the trials up to the fifth conditioned response [CR]), which is the stage thought to be most dependent on hippocampal activation. In addition, the use of yoked control groups in the current study allowed us to determine whether the presence of theta before a training trial enhanced learning rate in this hippocampus-dependent task, and/or whether nontheta is detrimental to learning as it is in delay conditioning. In addition to observing behavioral differences between the theta- and nonthetatriggered groups, we were interested in assessing whether theta triggering accelerates learning rate and hippocampal unit firing rate in this more difficult, hippocampus-dependent task (Moyer, Deyo, & Disterhoft, 1990; Solomon, Van der Schaaf, Thompson, & Weisz, 1986). To date, no studies have explored the relationship between theta-contingent trial presentation and changes in hippocampal unit responses. We expected to see robust hippocampal unit responses developing earlier in the theta-triggered group than in the nontheta-triggered group, especially during the trace period of the trial (see McEchron & Disterhoft, 1999). Such a demonstration would provide a direct link between the presence of theta, changes in the excitability of hippocampal neurons, and learning rate. Characterization of these relationships would establish a foundation and methodology for future investigations of a possible connection between hippocampal oscillatory potentials and plasticity throughout the essential and modulatory neural substrates of eyeblink conditioning. Subjects Method Subjects were 18 New Zealand White rabbits (Oryctolagus cuniculus) supplied by Myrtle s Rabbitry (Thompson Station, TN). All rabbits were maintained on a 12-hr light dark cycle, with training conducted during the light phase. The rabbits were allowed free access to food and water in their home cages. All procedures involving animals were approved by the Miami University Institutional Animal Care and Use Committee. Electrode Implantation All rabbits were anesthetized with ketamine (50 mg/kg im) and xylazine (10 mg/kg im) and implanted with bilateral hippocampal electrodes (size 00 insect pins insulated with Epoxylite [Epoxylite Corporation, Westerville, OH], except 50 70 m at the tip). The electrodes were positioned according to stereotaxic coordinates (Girgis & Shih-Chang, 1981; 4.5 mm posterior to bregma, 5.5 mm lateral to the midline suture and approximately 3.0 mm ventral to dura) and by monitoring activity from the electrode tip during implantation. Training After 5 days of postsurgical recovery and one 30 45-min session of adaptation to the restraint apparatus and conditioning chamber, rabbits began trace eyeblink conditioning. The paradigm, trace 500 (L. T. Thompson, Moyer, & Disterhoft, 1996) consisted of a 100-ms, 1kHz, 80 db tone followed by a 500-ms trace interval and a 3-psi, 100-ms corneal airpuff. Procedures for theta-contingent trial presentation have been described previously (Seager et al., 2002). Briefly, neural activity from one of the hippocampal electrodes was filtered (0.5 22.0 Hz) and then monitored in real time by a software program (Labview, National Instruments Corporation, Austin, TX) designed to compute a spectral ratio of the proportion of theta (3.5 8.5 Hz) to nontheta (0.5 3.5 Hz and 8.5 22.0 Hz) for 640-ms scrolling time intervals, updated every 160 ms. For the thetatriggered group (T ), trials were given only when the spectral ratio exceeded 1.0 three times in a row (960 ms total pretrial duration). For the nontheta-triggered (T ) group, trials were given only when the spectral ratio fell below 0.3 three times in a row. These criteria for T and T trials were developed during pilot experiments designed to maximize (or minimize) the probability that theta would continue throughout the conditioning trial. Yoked controls (Y and Y ; matched to the T and T rabbits by age and sex) were given the same ITI and trials per day as their T /T counterparts irrespective of the amount of theta occurring before each trial. Each session lasted for 90 min. Rabbits were trained until they reached a behavioral criterion of eight CRs out of nine consecutive trials, which is conventionally thought to be the point of asymptotic responding (Gormezano, Prokasy, & Thompson, 1987). We also used a second learning criterion, the number of trials to the fifth CR, which is indicative of initial acquisition of the CS US contingency (Prokasy, 1972, 1987; R. F. Thompson, Berry, Rinaldi, & Berger, 1979). Hippocampal Unit Analysis Multiple-unit activity from the electrodes was band-pass filtered (500 5000 Hz, Krohn-Hite Model 3700 filter [Krohn-Hite Corporation, Brockton, MA]) and passed through a window discriminator (DataWave Technologies, Broomfield, CO), which separated the largest spikes in the unit activity from background activity (approximately 2.5:1 signal-to-noise ratio). A computer sampled at the rate of 20 khz, calculated the number of spikes in each 10-ms bin for 100 bins and constructed daily average peristimulus time histograms. The hippocampal neural activity in response to conditioning stimuli was quantified by computing standard scores from each rabbit s daily histogram. Standard scores were computed by subtracting the average bin height of the pre-cs period from the height of each of the 100 bins and dividing by the standard deviation of the pre-cs period. Each histogram was divided into four periods, with each period summarizing neural responses during a portion of the training trial as follows: Period 1, tone CS (100 ms, 10 bins); Period 2, stimulus-free trace period (500 ms, 50 bins); Period 3, air US (100 ms, 10 bins); and Period 4, end of trial (300 ms, 30 bins). A total score was calculated for each period for each rabbit by adding together the individual standard score values for the appropriate bins. To test differences between groups, a 2 3 mixed design analysis of variance was used for each period of the trial in which group was a between-subjects variable and day was a within-subjects variable. Histology At the end of the experiment, rabbits were lightly anesthetized and a small marking lesion was made by passing a 200 A, 10-s direct current through each recording electrode (Grass Stimulator Model SD-9, Grass Instruments, West Warwick, RI). Rabbits were then given an overdose of sodium pentobarbital (Euthasol, 0.2205 mg/kg iv) and perfused intracardially with saline (0.9%) and Formalin (10%) solutions. The brains were removed, sectioned with a cryostat, embedded on gelatin-coated slides, stained with Prussian blue to mark the locations of the electrode tips, and counterstained with Safranin (Sigma, St. Louis, MO). Slides were examined with a compound microscope (Nikon, Japan) for verification of electrode locations. Only rabbits with electrodes in CA1 (stratum oriens or stratum pyramidale) were included in the study.

THETA-RELATED ACCELERATION OF LEARNING RATE 405 Results T Trial Presentation Examples of hippocampal slow-wave activity that triggered trials received by rabbits in the theta-contingent (T ) and nontheta-contingent (T ) groups are shown in Figure 1. It is clear that the recording from the T subject (see Figure 1A) is dominated by sinusoidal theta activity at approximately 5 Hz, whereas the recording from the T subject is less rhythmic, with substantial nontheta activity (see Figure 1B). Intermediate waveforms (see Method section) would not have triggered trials in either the T or T groups but were often seen during trials in the yoked controls. Behavioral Results As expected, rabbits that received training trials only while exhibiting theta activity (T ) showed an accelerated learning rate compared with those that received training trials in the explicit absence of theta activity (T ). More specifically, we expected to observe faster learning rates in the T group for both the early (fifth CR) learning criterion and the asymptotic (8/9) criterion. Figure 2 shows the percentage of CRs for the first 4 days of Figure 2. Mean ( SEM) percentage of conditioned responses (CRs) across the first 4 days of conditioning for the theta-contingent (T ) and nontheta-contingent (T ) groups. The T group gave a greater percentage of CRs across the first 4 days of conditioning than the T group ( p.01). training for the T and T groups. It is clear that the T group gave a significantly greater percentage of CRs than the T group over the first four sessions of conditioning. This observation was verified by a 2 4 (Group Day) mixed design analysis of variance, which showed a significant main effect of group, F(1, 7) 6.99, p.03. Figure 3 shows that, for the early stage (fifth CR) criterion, the T group (M 50.60, SD 21.18) learned significantly faster than control rabbits yoked to the T group (M 181.40, SD 140.47), t(4) 2.33, p.04, and 4 times faster than the T group (M 212.00, SD 75.79), t(7) 4.61, p.01. Similarly, the T group learned significantly more slowly than the T yoked control group (M 141.25, SD 53.24), t(4) 2.87, p.03. Figure 4 shows that for the asymptotic (8/9) learning criterion, the T group (M 180.20, SD 128.70) learned significantly faster than yoked controls (M 282.20, SD 187.04), t(4) 3.42, p.01, and the T group (M 399.25, SD 167.88), t(7) 2.22, p.03. The T group did not differ significantly from their yoked controls (M 363.25, SD 219.25), t(3) 0.33, p.39. As expected, because theta did not vary systematically between the two yoked control groups, these groups did not differ significantly from each other in learning rate for the fifth CR criterion, t(7) 0.54, p.61, or the 8/9 criterion, t(7) 0.60, p.57. Hippocampal Unit Activity Figure 1. Examples of hippocampal slow-wave activity that triggered trials in the theta-contingent (T ) and nontheta-contingent (T ) groups. A: Notice that slow-wave activity during T trials showed a predominance of activity in the theta band. B: Conversely, hippocampal activity during T trials showed a mixture of frequencies higher and lower than the theta band. In addition to behavioral differences between the T and T groups, we expected to see rapidly developing and more dramatic increases in hippocampal unit activity in the T group compared with the T group. Most of the T rabbits (4 out of 6) began giving CRs on or before the 3rd day of conditioning. In contrast, none of the T rabbits had begun giving CRs at this point. Therefore, we compared hippocampal firing rate in response to the conditioning stimuli between the two groups over the first 3 days of conditioning. As shown in Figures 5 and 6, the groups had

406 GRIFFIN, ASAKA, DARLING, AND BERRY Figure 3. Theta-contingent (T ) trial presentation reduces the number of trials required to establish the conditioned stimulus contingency (fifth conditioned response [CR]) in the early stage of learning (Stage 1). A: The T group (n 5) reached the fifth CR behavioral criterion more than 4 times faster than the nonthetacontingent (T ) group (n 4). ** p.01. B: The T group took significantly fewer trials to reach the fifth CR behavioral criterion than yoked controls. Conversely, the T group took significantly more trials to reach the fifth CR criterion than their yoked control group. *p.05, significantly different from zero. similar unit responses on Day 1, but the T group showed a significantly greater increase in hippocampal unit firing rate across days of conditioning than did the T group, which developed inhibitory responses. It is important that this increase occurred in the trace period of the trial, the period during which the hippocampus is thought to be most essential. Histograms of hippocampal unit activity show that the T and T groups both exhibited accelerations in unit firing rate during the US portion of the trial (see Figure 5A and 5B). However, this conditioning-related activity began to move to the trace interval of the trial as early as the 2nd day of conditioning in the T group, but was not evident in the T group even on the 3rd day of conditioning. In order to quantify and perform statistical verification of this observation, we computed standard scores of hippocampal unit activity in response to the training stimuli and compared these values between the T and T groups across the first 3 days of conditioning (see Figure 6). For the 100-ms tone period of the trial, there was a significant main effect of group, F(1, 13) 15.61, p.01, showing that the T group had significantly higher unit standard scores than the T group in response to the tone CS. For the 500-ms trace period of the trial, there was a significant Group Day interaction, F(2, 26) 5.77, p.01, with simple main effects tests revealing that Figure 4. Theta-contingent (T ) trial presentation reduces the number of trials required to reach asymptotic responding (eight of nine CRs). A: The T group (n 5) reached the 8/9 behavioral criterion more than twice as fast as the nontheta-contingent (T ) group, ** p.01. The T group took significantly fewer trials to reach the 8/9 behavioral criterion than yoked controls. B: However, the T group reached criterion only slightly (and not significantly) more slowly than their yoked controls, *p.05, significantly different from zero.

THETA-RELATED ACCELERATION OF LEARNING RATE 407 Figure 5. The theta-contingent (T ) group displayed a greater increase in hippocampal firing rate in the trace interval of the trial than the nontheta-contingent (T ) group. Standard scores of hippocampal unit firing rate across the first days of conditioning in a representative rabbit from the T group (left) and the T group (right). The arrows indicate tone (conditioned stimulus) and air (unconditioned stimulus) onset. the T group had significantly higher unit standard scores than the T group on Days 2, t(13) 2.25, p.04, and 3, t(13) 3.95, p.01, but not Day 1. The T group showed significant inhibitory responses to the tone, t(6) 12.52, p.01, and the trace period, t(6) 2.03, p.04, on Day 2. Similar results were seen on Day 3, with significant inhibitory responses to the tone periods, t(6) 10.21, p.01, and trace periods, t(6) 7.19, p.01. There were no group differences (main effects or Group Day interactions) for the air US portion of the trial or for the 300 ms following US offset. Discussion Together, the pattern of results in the current study suggests that the presence of theta activity during each conditioning trial enhances both the establishment of the CS US contingency, as revealed by the fifth CR criterion, and the length of time to reach stable, asymptotic responding (8/9 CRs). Conversely, triggering training trials when the hippocampal activity was explicitly low in theta significantly delayed only the initial acquisition of the CS US contingency, although asymptotic responding developed as quickly as in controls. In parallel with the enhancement of acquisition rate, theta-contingent trial presentation also facilitated the hippocampal conditioned unit response in the T as compared with the T group, suggesting that the presence of theta activity in the hippocampus may lead to increased plasticity of hippocampal circuitry. The demonstration that the two yoked control groups were similar in learning rate verifies that the acquisition rate differences

408 GRIFFIN, ASAKA, DARLING, AND BERRY Figure 6. Standard scores of hippocampal unit activity across the first 3 days of training for (A) the 100-ms tone period of the trial and (B) the 500-ms stimulus-free trace period. The theta-contingent (T ) group showed more dramatic unit responses than the nontheta-contingent (T ) group during both portions of the trial, especially on Days 2 and 3 of training. In fact, the T group showed significant inhibitory responses during the tone and trace periods. Because we quantified the unit responses by adding together the standard scores in each period, the difference in scale between Panels A and B is indicative of the difference in period duration. between the T and T groups were due to our theta manipulation, and not due to variations in ITI length and number of trials per day, which are known to affect learning rate (Gormezano, 1966). In fact, the yoked control group learning rates were intermediate between those of the T and T groups, as would be expected if these rabbits received some trials in theta and some trials in nontheta. Post hoc analyses verified this mixture of theta and nontheta trials in both control groups. These yoked control results are essentially identical to those obtained for the yoked control groups in the Seager et al. (2002) delay conditioning study. In that study, there was also a highly significant difference between T and T treatments; however, the magnitude of the T difference (from controls) was greater than the T difference. The highly significant T versus control difference in the current study may reflect the more essential hippocampal role in processes underlying trace versus delay conditioning. An intriguing possibility is that theta is beneficial to many classical conditioning tasks, whereas nontheta is especially detrimental to some (e.g., delay), a dissociation that may suggest ways to explore the relationship of theta to cognitive processes such as attention, awareness, and incentive motivation that can modulate learning rates. The major distinction between the behavioral results in the previous study using the delay paradigm (Seager et al., 2002) and the current study is that there was not a significant difference between the T and T groups in the number of trials to reach the fifth CR. This measure is thought to reflect the early, contingencydetection phase of learning (see Prokasy, 1972). The observation that the theta-contingent trial presentation leads to a dramatic initial learning enhancement in the trace, but not delay, paradigm suggests that theta activity plays a different role in hippocampusdependent tasks than it does in tasks for which its influence is primarily modulatory. Specifically, during trace conditioning, it is more important for hippocampal theta to be present on early trials, where theta may facilitate the formation of an initial association between the conditioning stimuli. Conversely, in the delay paradigm, the presence of hippocampal theta activity may be more important in establishing stable responding. The results of the present study serve as a further verification of the relationship between hippocampal theta activity and learning rate in a task that is known to depend on an intact hippocampal formation. Although the existence of a correlation has been known for many years, our recent development of a theta-contingent trial presentation protocol has allowed us to directly demonstrate that the presence of theta activity immediately preceding and continuing into the training trials leads to faster acquisition of the eyeblink response. This procedure has added support to the hypothesis, derived from earlier correlational and invasive studies, that naturally occurring theta activity plays a facilitory role in the establishment of a neural representation of the CS US contingency. Other findings support the notion that hippocampal theta may coordinate cellular activity and plasticity in areas outside the hippocampus proper (Dickson, Kirk, Oddie, & Bland, 1995; Frank, Brown, & Wilson, 2001; Vertes, Albo, Viana Di Prisco, 2001), suggesting more widespread hippocampal modulatory influences. Although previous studies have demonstrated that hippocampal neurons exhibit conditioning-related activity (see Berger, Rinaldi, Weisz, & Thompson, 1983; Berger & Thompson, 1978; Oliver, Swain, & Berry, 1993), this is the first investigation to compare conditioned unit responses between theta- and nontheta-triggered treatments. Our demonstration that the T group showed a greater acceleration of hippocampal firing rate early in conditioning strengthens the idea that the presence of theta activity is associated with greater plasticity of hippocampal neurons. However, the group difference did not emerge until Day 2 of conditioning, suggesting that theta enables a slowly developing plasticity in the hippocampus that may allow the subject to learn faster. Moreover, the observation that the difference between T and T groups in firing rate occurred during the trace interval of the trial is consistent with the assertion that the trace paradigm requires an intact hippocampus, and suggests that the role of the hippocampus is to bridge the time interval between the discontiguous events of the trial (see Wallenstein et al., 1998). Our results are consistent with

THETA-RELATED ACCELERATION OF LEARNING RATE 409 a previous study (McEchron & Disterhoft, 1999) that examined hippocampal single-unit activity during trace rabbit eyeblink conditioning and found that learning-related acceleration of unit firing rate first appeared in response to the US, and subsequently began to occur earlier in the trial (in the trace interval). This pattern of unit activity was then followed by behavioral learning, suggesting that the hippocampal conditioned unit response occurs prior to the appearance of stable conditioned responses, and more importantly, that the increase in firing rate must occur in the trace interval of the trial in order for behavioral learning to take place. Because multiple-unit recordings were used to index hippocampal firing rate, it is impossible to determine whether the group differences reflect direct inhibition of hippocampal pyramidal cells or reduced excitation. In fact, McEchron and Disterhoft (1999) reported that single units show heterogeneous responses and that some neurons actually show inhibitory responses after learning has reached asymptotic levels. Future studies could address this issue by using spike-sorting techniques to isolate single neurons and thus determine the firing properties of individual hippocampal neurons in theta- and nontheta-triggered animals. Although demonstrations of learning impairments due to the disruption of pathways and structures that support the theta rhythm are numerous, less common are studies that report facilitation of learning due to increases in theta activity. One such study (Berry & Swain, 1989) found that water restriction caused an increase in theta activity and, consequently, in learning rate. It is important that this enhanced learning rate was accompanied by increased conditioned unit responses to training stimuli, which reflect rapidly developing hippocampal plasticity in the eyeblink paradigm (Berger et al., 1983). The water restriction study is a clear demonstration that the motivational state of the animal (i.e., thirst) can affect theta activity and thereby lead to changes in the plasticity of hippocampal output neurons. It is by such a mechanism that the hippocampus is thought to play a modulatory role in associative learning through interaction with other brain systems, especially the cerebellum (Clark, McCormick, Lavond, & Thompson, 1984). Because theta was measured pretraining, the prior study could only infer that the water-deprived animals had more theta during the conditioning session itself. By using our theta-triggering technique, we were able to ensure that trials were received in the presence or absence of theta, making a stronger case for theta s enhancement of conditioned unit activity and learning. There have been several additional lines of research suggesting that theta leads to an increase in hippocampal plasticity. Long-term potentiation is maximally induced at theta frequency and during periods of theta activity, suggesting a role for theta if, in fact, long-term potentiation underlies hippocampal unit responses and behavioral learning (Holscher, Anwyl, & Rowan, 1997; Larson, Wong, & Lynch, 1986; Pavlides, Greenstein, Gudman, & Winson, 1988; Thomas, Watabe, Moody, Makhinson, & O Dell, 1998). Relationships among theta, learning, and conditioned hippocampal unit responses have also been shown in a fear-conditioning paradigm in rats (Maren, DeCola, Swain, Fanselow, & Thompson, 1994). In summary, a growing literature suggests that the presence of high levels of theta may be optimal for learning as a result of its proposed facilitation of hippocampal cellular plasticity, which subsequently leads to modulation of the circuits necessary for the acquisition and execution of the learned behavioral response. In simpler paradigms that do not absolutely require an intact hippocampal formation, such as delay eyeblink conditioning (Akase, Alkon, & Disterhoft, 1989; Schmaltz & Theios, 1972; Solomon & Moore, 1975), this modulation has a facilitory but nonessential role (Berry & Seager, 2001). As in the current study, paradigms such as trace eyeblink conditioning that require hippocampal circuitry demonstrate an even stronger theta learning relationship. In fact, using the presence or absence of theta activity as an independent variable allows us to conclude that theta activity produces, or is simultaneous with, optimal conditions for learning. Perhaps the presence of theta is indicative of increased attention or awareness of environmental stimuli, leading to better performance, especially during difficult tasks such as trace conditioning. A recent study has investigated the relationship between learning rate in human eyeblink conditioning and the extent to which the participants were aware of the stimulus contingencies and found that this reported awareness enhanced learning rate (Manns, Clark, & Squire, 2000). Our findings lend empirical support to recent theories and computational models claiming a relationship between neurobiological oscillatory potentials and learning-related plasticity (Buszaki, 2002; Singer, 1993; Hasselmo, Bodelon, & Wyble, 2002) and may strengthen other models of hippocampal function (Gluck & Meyers, 1997; Meyers et al., 1996; Schmajuk & DiCarlo, 1992). An important next step is to investigate whether thetacontingent trial presentation is able to attenuate learning impairments. Recent laboratory results indicate that age-related learning impairments can be ameliorated by triggering training trials during theta activity, with aging animals in the T group learning as quickly as their younger counterparts (Asaka & Berry, 2004). Another exciting possibility is the application of the thetacontingent paradigm to human learning. There is growing evidence that theta activity is present in human subjects during the performance of cognitive tasks such as virtual maze navigation (Caplan, Madsen, Raghavachari, & Kahana, 2001), working memory (Raghavachari et al., 2001; Tesche & Karhu, 2000), and verbal behavior (Arnolds et al., 1980; Meador et al., 1991). Whether such learning can be optimized or deficits overcome by theta triggering is a matter for future research. In addition, extending thetacontingent training to human eyeblink classical conditioning would facilitate comparisons to known hippocampal and cerebellar substrates of human memory and their impairment by neurological disorders such as Alzheimer s disease (Woodruff-Pak, 1999). Finally, use of this methodology while recording simultaneous cellular activity in hippocampal efferent targets will begin to address the fundamental question of how the slow-wave state of the hippocampus affects other learning-related circuits to predict and modulate behavioral learning. References Adey, W. R. (1966). Neurophysiological correlates of information transaction and storage in brain tissue. In E. Stellar & J. M. Sprague (Eds.), Progress in physiological psychology (Vol. 1, pp. 1 43). New York: Academic Press. Akase, E., Alkon, D. L., & Disterhoft, J. F. (1989). Hippocampal lesions impair memory of short-delay conditioned eyeblink in rabbits. Behavioral Neuroscience, 103, 935 943. Arnolds, D. E., Lopes da Silva, F., Aitink, J. W., Kamp, A., & Boeijinga, P. (1980). The spectral properties of hippocampal EEG related to behaviour in man. Electroencephalography and Clinical Neurology, 50, 324 328.

410 GRIFFIN, ASAKA, DARLING, AND BERRY Asaka, Y., & Berry, S. D. (2004). [Age-related learning deficits: Reversal by a hippocampal theta-dependent brain computer interface]. Unpublished raw data. Asaka, Y., Griffin, A. L., & Berry, S. D. (2002). Reversible septal inactivation disrupts hippocampal slow-wave and unit activity and impairs trace conditioning in rabbits (Oryctolagus cuniculus). Behavioral Neuroscience, 116, 434 442. Asaka, Y., Seager, M. A., Griffin, A. L., & Berry, S. D. (2000). Medial septal microinfusion of scopolamine disrupts hippocampal activity and trace jaw movement conditioning. Behavioral Neuroscience, 114, 1068 1077. Bennett, T. L. (1969). The electrical activity of the hippocampus and processes of attention. In R. L. Isaacson & K. H. Pribram (Eds.), The hippocampus (Vol. 2, pp. 71 99). New York: Plenum Press. Berger, T., Rinaldi, P., Weisz, D., & Thompson, R. F. (1983). Single-unit analysis of different hippocampal cell types during classical conditioning of the rabbit nictitating membrane response. Journal of Neurophysiology, 50, 1197 1219. Berger, T. W., & Thompson, R. F. (1978). Neuronal plasticity in the limbic system during classical conditioning of the rabbit nictitating membrane response: I. The hippocampus. Brain Research, 145, 323 346. Berry, S. D., & Seager, M. A. (2001). Hippocampal theta oscillations and classical conditioning. Neurobiology of Learning and Memory, 76, 298 313. Berry, S. D., Seager, M. A., Asaka, Y., & Borgnis, R. A. (2000). Motivational issues in aversive and appetitive conditioning paradigms. In D. S. Woodruff-Pak & J. E. Steinmetz (Eds.), Classical conditioning (Vol. 2, pp. 287 312). Boston: Kluwer Academic. Berry, S. D., & Swain, R. A. (1989). Water deprivation optimizes hippocampal activity and facilitates nictitating membrane conditioning. Behavioral Neuroscience, 103, 71 76. Berry, S. D., & Thompson, R. F. (1978, June 16). Prediction of learning rate from the hippocampal electroencephalogram. Science, 200, 1298 1300. Berry, S. D., & Thompson, R. F. (1979, July 13). Medial septal lesions retard classical conditioning of the nictitating membrane response in rabbits. Science, 205, 209 211. Buszaki, G. (2002). Theta oscillations in the hippocampus. Neuron, 33, 325 340. Caplan, J. B., Madsen, J. R., Raghavachari, S., & Kahana, M. J. (2001). Distinct patterns of brain oscillations underlie two basic parameters of human maze learning. Journal of Neurophysiology, 86, 368 380. Clark, G. A., McCormick, D. A., Lavond, D. G., & Thompson, R. F. (1984). Effects of lesions of cerebellar nuclei on conditioned behavioral and hippocampal neuronal responses. Brain Research, 291, 125 136. Dickson, C. T., Kirk, I. J., Oddie, S. D., & Bland, B. H. (1995). Classification of theta-related cells in the entorhinal cortex: Cell discharges are controlled by the ascending brainstem synchronizing pathway in parallel with hippocampal theta-related cells. Hippocampus, 5, 306 319. Eichenbaum, H., Otto, T., & Cohen, N. J. (1992). The hippocampus: What does it do? Behavioral and Neural Biology, 57, 2 36. Frank, L. M., Brown, E. N., & Wilson, M. A. (2001). A comparison of the firing properties of putative excitatory and inhibitory neurons from CA1 and the entorhinal cortex. Journal of Neurophysiology, 86, 2029 2040. Girgis, M., & Shih-Chang, W. (1981). Stereotaxic atlas of the rabbit brain. St. Louis: Warren H. Green. Givens, B. S., & Olton, D. S. (1990). Cholinergic and GABAergic modulation of medial septal area: Effect on working memory. Behavioral Neuroscience, 104, 849 855. Gluck, M. A., & Meyers, C. E. (1997). Psychobiological models of hippocampal function in learning and memory. Annual Review of Psychology, 48, 481 514. Gormezano, I. (1966). Classical conditioning. In J. B. Sidowski (Ed.), Experimental methods and instrumentation in psychology (pp. 385 420). New York: McGraw-Hill. Gormezano, I., Prokasy, W. F., & Thompson, R. F. (Eds.). (1987). Classical conditioning (3rd ed.). Hillsdale, NJ: Erlbaum. Green, J. D., & Arduini, A. (1954). Hippocampal electrical activity in arousal. Journal of Neurophysiology, 17, 533 557. Hasselmo, M. E., Bodelon, C., & Wyble, B. P. (2002). A proposed function for hippocampal theta rhythm: Separate phases of encoding and retrieval enhance reversal of prior learning. Neural Computation, 14, 793 817. Holscher, C., Anwyl, R., & Rowan, M. J. (1997). Stimulation on the positive phase of hippocampal theta rhythm induces long-term potentiation that can be depotentiated by stimulation on the negative phase in area CA1 in vivo. Journal of Neuroscience, 17, 6470 6477. Kahana, M. J., Seelig, D., & Madsen, J. R. (2001). Theta returns. Current Opinion in Neurobiology, 11, 739 744. Kaneko, T., & Thompson, R. F. (1997). Disruption of trace conditioning of the nictitating membrane response in rabbits by central cholinergic blockade. Psychopharmacology, 131, 161 166. Larson, J., Wong, D., & Lynch, G. (1986). Patterned stimulation at the theta frequency is optimal for the induction of hippocampal long-term potentiation. Brain Research, 368, 347 350. Lawson, V. H., & Bland, B. H. (1993). The role of the septohippocampal pathway in the regulation of hippocampal field activity and behavior: Analysis by the intraseptal microinfusion of carbachol, atropine, and procaine. Experimental Neurology, 120, 132 144. Manns, J. R., Clark, R. E., & Squire, L. R. (2000). Awareness predicts the magnitude of single-cue trace eyeblink classical conditioning. Hippocampus, 19, 181 186. Maren, S., DeCola, J. P., Swain, R. A., Fanselow, M. S., & Thompson, R. F. (1994). Parallel augmentation of hippocampal long-term potentiation, theta rhythm, and contextual fear conditioning in water-deprived rats. Behavioral Neuroscience, 108, 44 56. Markowska, A., Olton, D., & Givens, B. (1995). Cholinergic manipulations in the medial septal area: Age-related effects on working memory and hippocampal electrophysiology. Journal of Neuroscience, 15, 2063 2073. McEchron, M. D., & Disterhoft, J. F. (1999). Hippocampal encoding of non-spatial trace conditioning. Hippocampus, 9, 385 396. Meador, K. J., Thompson, J. L., Loring, D. W., Murro, A. M., King, D. W., Gallagher, B. B., et al. (1991). Behavioral state-specific changes in human hippocampal theta activity. Neurology, 41, 869 872. Mizumori, S. J. Y., Perez, G. M., Alvarado, M. C., Barnes, C. A., & McNaughton, B. L. (1990). Reversible inactivation of the medial septum differentially affects two forms of learning in rats. Brain Research, 528, 12 20. Moyer, J., Deyo, R., Disterhoft, J. F. (1990). Hippocampectomy disrupts trace eye-blink conditioning in rabbits. Behavioral Neuroscience, 104, 243 252. Myers, C. E., Ermita, B. R., Harris, K., Hasslemo, M., Solomon, P., & Gluck, M. A. (1996). A computational model of cholinergic disruption of septohippocampal activity in classical eyeblink conditioning. Neurobiology of Learning and Memory, 66, 51 66. Oliver, C. G., Swain, R. A., & Berry, S. D. (1993). Hippocampal plasticity during jaw movement conditioning in the rabbit. Brain Research, 608, 150 154. Pavlides, C., Greenstein, Y. J., Gudman, M., & Winson, J. (1988). Longterm potentiation in the dentate gyrus is induced preferentially on the positive phase of theta rhythm. Brain Research, 439, 383 387. Prokasy, W. F. (1972). Developments with the two-phase model applied to human eyelid conditioning. In A. H. Black & W. F. Prokasy (Eds.), Classical conditioning: II. Current research and theory (pp. 119 147). New York: Appleton-Century-Crofts. Prokasy, W. F. (1987). A perspective on the acquisition of skeletal responses employing the Pavlovian paradigm. In I. Gormezano, W. F.

THETA-RELATED ACCELERATION OF LEARNING RATE 411 Prokasy, & R. F. Thompson (Eds.), Classical conditioning (3 rd ed., pp. 287 318). Hillsdale, NJ: Erlbaum. Raghavachari, S., Kahana, M. J., Rizzuto, D. S., Caplan, J. B., Kirschen, M. P., Bourgeois, B., et al. (2001). Gating of human theta oscillations by a working memory task. Journal of Neuroscience, 21, 3175 3183. Redish, A. D. (2001). The hippocampal debate: Are we asking the right questions? Behavioural Brain Research, 127, 81 98. Salvatierra, A. T., & Berry, S. D. (1989). Scopolamine disruption of septo-hippocampal activity and classical conditioning. Behavioral Neuroscience, 103, 715 721. Schmajuk, N. A., & DiCarlo, J. J. (1992). Stimulus configuration, classical conditioning, and hippocampal function. Psychological Review, 99, 268 305. Schmaltz, L. W., & Theios, J. (1972). Acquisition and extinction of a classically conditioned response in hippocampectomized rabbits (Oryctolagus cuniculus). Journal of Comparative and Physiological Psychology, 79, 328 333. Seager, M. A., Johnson, L. D., Chabot, E. S., Asaka, Y., & Berry, S. D. (2002). Oscillatory brain states and learning: Impact of hippocampal theta-contingent training. Proceedings of the National Academy of Sciences, USA, 99, 1616 1620. Singer, W. (1993). Synchronization of cortical activity and its putative role in information processing and learning. Annual Review of Physiology, 55, 349 374. Solomon, P. R., & Moore, J. W. (1975). Latent inhibition and stimulus generalization of the classically conditioned nictitating membrane response in rabbits (Oryctolagus cuniculus) following dorsal hippocampal ablation. Journal of Comparative and Physiological Psychology, 89, 1192 1203. Solomon, P. R., Van der Schaaf, E., Thompson, R. F., & Weisz, D. (1986). Hippocampus and trace conditioning of the rabbit s classically conditioned nictitating membrane response. Behavioral Neuroscience, 100, 729 744. Tesche, C. D., & Karhu, J. (2000). Theta oscillations index human hippocampal activation during a working memory task. Proceedings of the National Academy of Sciences, USA, 97, 919 924. Thomas, M. J., Watabe, A. M., Moody, T. D., Makhinson, M., & O Dell, T. J. (1998). Postsynaptic complex spike bursting enables the induction of LTP by theta frequency synaptic stimulation. Journal of Neuroscience, 18, 7118 7126. Thompson, L. T., Moyer, J. R., & Disterhoft, J. F. (1996). Trace eyeblink conditioning in rabbits demonstrates heterogeneity of learning ability both between and within age groups. Neurobiology of Aging, 17, 619 629. Thompson, R. F., Berry, S. D., Rinaldi, P. C., & Berger, T. W. (1979). Habituation and the orienting reflex: The dual process revisited. In H. D. Kimmel, E. H. Van Olst, & J. F. Orlebeke (Eds.), The orienting response in human (pp. 21 60). Hillsdale, NJ: Erlbaum. Tracy, A. L., Jarrard, L. E., & Davidson, T. L. (2001). The hippocampus and motivation revisited: Appetite and activity. Behavioural Brain Research, 127, 13 23. Vanderwolf, C. H. (1971). Limbic-diencephalic mechanisms of voluntary movement. Psychological Review, 78, 83 113. Vertes, R. P., Albo, Z., & Viana Di Prisco, G. (2001). Theta-rhythmically firing neurons in the anterior thalamus: Implications for mnemonic functions of Papez s circuit. Neuroscience, 104, 619 625. Wallenstein, G. V., Eichenbaum, H. B., & Hasselmo, M. E. (1998). The hippocampus as an associator of discontiguous events. Trends in Neurosciences, 21, 317 323. Winson, J. (1978, July 14). Loss of hippocampal theta rhythm results in spatial memory deficit in the rat. Science, 210, 160 163. Woodruff-Pak, D. S. (1999). New directions for a classical paradigm: Human eyeblink conditioning. Psychological Science, 10, 1 3. Received May 15, 2003 Revision received September 17, 2003 Accepted September 29, 2003