Schedule control of behavior reinforced by electrical stimulation of the brain.

Electrical stimulation of the brain was used to train rats to respond on random interval schedules. Stimulation was either delayed for 0.5 second and preceded by a brief signal, delayed and unsignaled, or presented contiguously with the response. In every case, responding was maintained on schedules and showed resistance to extinction typical of food-reinforced responding. Priming was never necessary. These data cast doubt on the generality of beliefs about the behavioral effects of brain stimulation reinforcement.

The discovery of self-stimulation and other stories.

The author's recollections of the events leading to the discovery of rewarding brain stimulation at McGill University in 1953, with a history of his subsequent attempts to find a learning theory congruent with the phenomenon.

A cell assembly theory of hippocampal amnesia.

Recent memories are more susceptible to amnesic loss than older memories, the time scale being much longer than can reasonably be explained by a passive chemical or morphological change. A possible explanation is that memories are initially sustained by "soft", easily produced but ephemoral, synaptic changes to which are later added "hard" changes that are more durable but require repeated synaptic activity over a long period to become established. "Soft" synapses are assumed to be concentrated in parts of the limbic system, "hard" synapses in the neocortex. The theory can also explain why objects encountered by patients with anterograde amnesia never become familiar to them.

Behavioral measurement of axonal thresholds.

A behavioral method for measuring the electrical sensitivity of directly stimulated elements in the brain is described and applied to the medial forebrain bundle (MFB) reward path and the tectospinal circling path. Equations are derived from which threshold current densities may be calculated from knowledge of the electrode tip dimensions and the current required to produce criterion behavior, which is a function of electrode size. Four different sizes of electrode were implanted in the MFB of rats and self-stimulation rates plotted against stimulating current. The mean currents for criterion bar-pressing rates of 25% and 55% of maximum rate were determined for each electrode size and the values used to calculate average threshold current densities. Two sizes of electrode were implanted in the tectospinal tract of rats and the average currents to produce circling at 0.2 and 0.4 turns/s were measured. The threshold current densities for self-stimulation axons were about 5 times as large as those for circling, in accordance with other evidence that tectal circling path axons are larger than those of the MFB reward path.

Short- and long-term summation characteristics of electrical self-stimulation reward.

Using a Y-maze preference test paradigm, we examined the characteristics of the neural networks that integrate trains of rewarding stimulation pulses. Rats compared the rewarding effectiveness of various durations of a test reinforcement to those of 3 durations of a standard reinforcement. By changing the duration of the standard reinforcement while keeping the pulse frequency constant, we varied the stimulation magnitude to which the test reinforcement was being compared. This enabled us to examine the relationships between stimulation parameters both within a constant stimulation magnitude and across different stimulation magnitudes. The data were inconsistent with a simple integrating system with a single decay time constant such as is usually assumed. They can most parsimoniously be accounted for by two systems having quite widely differing time constants. The time constant of the first integrator is ca. 450 ms, whereas that of a second integrator is ca. 6.5 s.

Fiber pathways associated with cerebellar self-stimulation in the rat: a retrograde and anterograde tracing study.

Intracranial self-stimulation (ICSS) was obtained from an area of cerebellum just rostro-ventral to the fastigial nucleus. The acquisition of cerebellar ICSS was slow, although this depended in part on the type of operant task required. The afferent and efferent fiber connections of the region of cerebellum supporting ICSS were identified using silver-degeneration and horseradish peroxidase tracing techniques. Two major pathways, one ascending to the ventromedial thalamus and one descending to the paramedian reticular formation and the region of the solitary nucleus, are discussed as possible substrates of ICSS at sites in the cerebellum as well as from other pontine and medullary regions.

The role of corticocortical projections in self-stimulation of the prelimbic and sulcal prefrontal cortex in rats.

Four experiments were performed to assess the nature of the contribution of the corticocortical projections between the prelimbic and sulcal divisions of the rat prefrontal cortex to self-stimulation (SS) of these sites. The first experiment showed that transection of these projections by parasagittal knife cuts or bilateral electrolytic lesions of the prelimbic cortex had no effect on SS of the sulcal cortex. The second experiment demonstrated that SS of the prelimbic cortex could be obtained after transection of the corticocortical projection path. The third experiment demonstrated that the deficit in prelimbic SS, seen to follow such bilateral transections, is a function of the amount of exposure to the stimulation given to the animals after the lesion. The fourth experiment showed that the stimulation-dependent process underlying the acquisition of prelimbic and sulcal SS could be dissociated by the knife cuts. The discussion focused on the implications of these findings for an account of prefrontal self-stimulation behavior.

Rat strain differences in the acquisition of hippocampal self-stimulation.

The rate of acquisition of lever-pressing for electrical stimulation of the hippocampus (HPC) was compared in two strains of rat: barrier-sustained Wistar and Sprague-Dawley. Sprague-Dawley rats initially bar-pressed at very low rates and took a median of 11 days to self-stimulate, according to the criterion used. Wistar rats all reached the same criterion in the first test session. Differential sensitivity to the activating effects of stimulation as an explanation for this difference was ruled out by the observation that both strains decreased response rates at the same rate and to the same level if stimulation was made non-contingent on lever-pressing. Differential threshold for reward was ruled out by the observation that rate-intensity curves yielded the same threshold currents and peak rates in both strains. Finally, it was shown that the rate of development of kindled seizures in the two strains of rats is different: Wistars kindle to full seizures faster than do Sprague-Dawleys. The relationship between the quicker onset of self-stimulation and of kindled seizures in Wistars is discussed.

Strength-duration characteristics of lateral hypothalamic and periaqueductal gray reward-path neurons.

Behaviorally measured strength-duration curves were plotted for reward related neurons stimulated by electrodes in the lateral hypothalamus (LH) and ventral periaqueductal gray (PG). The average curves for the two groups were only marginally different, suggesting that similar neural populations mediate self-stimulation at the two sites. Analysis of the data for both groups indicates that stimulating pulses shorter than 0.1 msec stimulate mainly a short time-constant substrate with an average time-constant of about 0.1 msec. This is the same time-constant that has been reported in the literature for myelinated axons having a wide range of conduction velocities. It is suggested that the marked departure of behaviorally determined strength-duration curves from the approximately exponential form usually exhibited by single elements is due to the stimulation of unmyelinated elements by long (but not by short) stimulation pulses.

Temporal characteristics of electrical self-stimulation reward: fatigue rather than adaptation.

Using a Y-maze preference test paradigm, we examined the temporal characteristics of the neural network subserving self-stimulation reward. Rats were given a choice between two pulse trains of stimulation, which varied in duration and pulse frequency. The results showed that increasing the pulse frequency decreases the duration at which the rewarding effectiveness of brain stimulation reaches an asymptote. The data also indicated that when prolonged stimulation is delivered at a high pulse frequency, the initial pulses contribute the most to the rewarding effect. Later pulses are affected by the reduced ability of the neurons or synapses to transmit signals along the neural network due to fatigue. The data are explained in terms of an improved model of summation involving more than one integrator and fatigue.

Further evidence against adaptation of prolonged electrical self-stimulation reward.

Using a Y-maze preference test paradigm, we examined the temporal characteristics of the neural network subserving self-stimulation reward. The first part of the experiment demonstrated that when prolonged electrical brain stimulation is initially delivered with a low pulse frequency (100 Hz), rats prefer an increase over either a decrease or no change in the pulse frequency of subsequent stimulation. However, the second part showed that when prolonged brain stimulation is initially delivered with a high pulse frequency (250 Hz), an increase is not preferred. The data are inconsistent with an adaptation model of summation. These results are explained in terms of an improved model of summation involving two integrators and fatigue.

Development of brain stimulation reward in the medial prefrontal cortex: facilitation by prior electrical stimulation of the sulcal prefrontal cortex.

Electrical stimulation of the medial prefrontal cortex (MC) in rats delivered daily for seven days causes a marked improvement in the rate of acquisition of a self-stimulation response. In the present experiment, we looked at whether we could get the same facilitatory effect on self-stimulation of the MC by delivering pre-training stimulation to other points in the brain anatomically related to the MC. Electrical stimulation of the lateral hypothalamus was without effect. However, electrical stimulation of the sulcal prefrontal cortex (SC) either contralateral or ipsilateral to the MC electrode did facilitate acquisition of self-stimulation of the MC. Thus the Sc and MC would appear to be part of the same substrate controlling the development of positive reinforcement in the MC.

Treatment with anticonvulsant drugs retards the development of brain-stimulation reward in the prefrontal cortex.

Electrical stimulation of the medial prefrontal cortex in rats daily for nine days caused a marked improvement in the rate of acquisition of a self-stimulation response. Diazepam (1 mg/kg, IP) or phenobarbital (15 mg/kg), but not phenytoin (25 mg/kg), administered during the nine day period of electrical stimulation, attenuated this facilitatory effect. However, diazepam or phenobarbital in the same dosages administered to self-stimulating rats (i.e., after acquisition) failed to alter responding. It was suggested that a kindling-like mechanism may underlie the development of self-stimulation of the prefrontal cortex.

Distinct substrates influence the acquisition of self-stimulation of the hippocampus and the prefrontal cortex.

Self-stimulation (SS) of both the medial prefrontal cortex (MPFC) and the dorsolateral hippocampus (HPC) is known to develop slowly, over a period of days. In both cases, the acquisition of bar-pressing can be markedly hastened by delivery of noncontingent electrical stimulation for several days prior to SS training. The similarity of these effects suggests that there might be a common substrate mediating the acquisition process. However, in the present experiment, pre-training noncontingent electrical stimulation of the MPFC had no effect on how rapidly rats acquired the bar-pressing response for HPC stimulation, or vice versa. A further dissociation of the elements governing the acquisition process for these two SS sites was suggested by the observation that pre-training noncontingent stimulation of the entorhinal cortex facilitated the speed of acquisition of SS of the HPC but not of the MPFC. It seems that the HPC and entorhinal cortex can be excluded from the subset of neural structures which are known to influence the acquisition process governing MPFC SS. These and other data suggest that the development of SS of the MPFC and HPC can be regarded, at least in part, as involving a process rooted in distinct substrates.

Elimination of medial prefrontal cortex self-stimulation following transection of efferents to the sulcal cortex in the rat.

Intracranial self-stimulation (ICSS) of the medial prefrontal cortex (MFC) was not affected by lesions of the medial forebrain bundle, the nucleus accumbens or medialis dorsalis. However, bilateral, parasagittal knife cuts that transected fibers interconnecting the medial and sulcal cortices eliminated ICSS from the MFC with no apparent recovery over a 21 day test period. Similar knife cuts produced only transient effects on lateral hypothalamic ICSS. These data suggest that the neural substrates of frontal cortex ICSS are very different than those that subserve ICSS along the medial forebrain bundle.

Plasticity of the medial prefrontal cortex: facilitated acquisition of intracranial self-stimulation by pretraining stimulation.

Prior electrical stimulation of the medial prefrontal cortex MFC facilitated the subsequent acquisition of intracranial self-stimulation (ICSS) from the same MFC electrode site. Stimulations that were spaced over a period of six days were more effective in producing this facilitation than the same number of stimulations delivered over a two day period. These data suggest that the rewarding effects of MFC stimulation may involve some process akin to the kindling phenomenon and as such may provide insights in the neuronal modifications thought to underlie learning and memory.

Stimulation parameters influence response involvement in reinforcing brain stimulation.

Rats pressed a lever for brain stimulation in the start box of a T-maze. Pulse trains of stimulation were available under various temporal schedules. Periods of self-stimulation (SS) alternated with intervals of experimenter-administered stimulation (EAS) during which identical stimulation was automatically delivered at the same average rate as during SS periods. Rats could terminate the on going EAS by traversing the maze and making a turn into the correct arm, thereby reinstating the availability of a SS period. Experiments 1 and 2 demonstrated that manipulations of either the intertrain interval or the train duration, which are parameters that regulate the temporal density of stimulation, influenced the latency to terminate intervals of EAS. Experiment 3 showed that omitting the stimulation during the EAS period of the standard SS-EAS cycle disrupted the resetting behaviour of the experienced rats. They required as many sessions as naive rats to learn the behaviour required to circumvent the EAS-free period. The data suggest that the behaviour patterns exhibited are dependent upon the buildup of activity that is largely influenced by the total neural activity summated across pulse train(s) of stimulation. The behaviour to terminate the EAS occurs as a consequence of an aversive effect which is concurrently produced by the rewarding stimulation. This aversive effect is attenuated by responding.

Effects of dopamine supersensitivity on lateral hypothalamic self-stimulation in rats.

Dopamine (DA) receptor supersensitivity was demonstrated by potentiated d-amphetamine stereotype after a three-day treatment regimen in which the DA receptor blocker pimozide (4.0 mg/kg) was administered twice daily. Similarly-induced DA supersensitivity produced a significant increase in the rate of lever-pressing for lateral hypothalamic (LH) intracranial self-stimulation (ICSS) and a significant decrease in ICSS thresholds. No change from pretreatment baselines was observed in vehicle-treated control animals. Following three-day treatment with the noradrenaline--(NA) and DA-receptor blocker, haloperidol (4.0 mg/kg twice daily), a single injection of the alpha-adrenergic agonist clonidine (0.15 mg/kg) caused increased running behavior. In contrast clonidine decreased running in rats pretreated with chronic pimozide or vehicle. These results indicate an increase in the sensitivity of central NA receptors following chronic haloperidol but not chronic pimozide. Taken together, these findings were interpreted as a potentiation in the reinforcing properties of LH-ICSS after chronic pimozide treatments due to increases in the sensitivity of DA and not NA receptors.

The mind and Donald O. Hebb.

By rooting behavior in ideas, and ideas in the brain, Hebb laid the groundwork for modern neuroscience. His theory prefigured computer models of neural networks.