Nature Protocols 2, 2987 - 2995 (2007)
Published online: 15 November 2007 | doi:10.1038/nprot.2007.441

Subject Categories: Model organisms | Neuroscience | Pharmacology and toxicology

Intracranial self-stimulation (ICSS) in rodents to study the neurobiology of motivation

William A Carlezon Jr1 & Elena H Chartoff1

It has become increasingly important to assess mood states in laboratory animals. Tests that reflect reward, reduced ability to experience reward (anhedonia) and aversion (dysphoria) are in high demand because many psychiatric conditions that are currently intractable in humans (e.g., major depression, bipolar disorder, addiction) are characterized by dysregulated motivation. Intracranial self-stimulation (ICSS) can be utilized in rodents (rats, mice) to understand how pharmacological or molecular manipulations affect the function of brain reward systems. Although many different methodologies are possible, we will describe in this protocol the use of medial forebrain bundle (MFB) stimulation together with the 'curve-shift' variant of analysis. This combination is particularly powerful because it produces a highly reliable behavioral output that enables clear distinctions between the treatment effects on motivation and the treatment effects on the capability to perform the task.



What is intracranial self-stimulation?

Intracranial self-stimulation (ICSS) is an operant paradigm in which rodents self-administer rewarding electrical stimulation (often referred to as brain stimulation reward (BSR)) through electrodes implanted into the brain. Though the terms are often used interchangeably, ICSS is the behavior and BSR is what is earned by the behavior. There are many permutations of the ICSS paradigm: rats will self-administer stimulation into many brain areas, the stimulation can be either monopolar or bipolar, and numerous approaches can be used to determine the amount of stimulation that is rewarding. An exhaustive description of all of these methodologies is well beyond the scope of this protocol. Rather, we will describe the methods we utilize, because they have proven useful for detecting how pharmacological or molecular manipulations affect motivation. We acknowledge that investigators who use ICSS (including us) tend to be extremely loyal to their preferred method.

ICSS is practical in both rats and mice. The most prominent difference between the procedures we use for rats and those we use for mice involves the manipulandum: we use a lever for rats, whereas we use a wheel for mice, to facilitate sustained responding in this species. In our experience, there are no detectable differences in ICSS behavior among mouse strains, including Swiss-Webster (Fig. 1) or C57BL/6 (see Supplementary Video 1 online). We implant the stimulating electrodes into the medial forebrain bundle (MFB) for all of our studies. Stimulation of this pathway produces reliable ICSS at relatively low current intensities, and is associated with few (if any) of the motor side effects that can occur with other electrode placements (e.g., ventral tegmental area (VTA))1. Presumably, MFB stimulation activates excitatory (possibly cholinergic) inputs to the mesolimbic dopamine system, thereby transsynaptically activating this brain reward pathway2, 3. We have shown that manipulations in projection regions of the VTA (e.g., the nucleus accumbens (NAc)) have profound effects on ICSS behavior4, 5, confirming that this pathway plays an important role in BSR.

Figure 1: Key elements of the operant chamber setup.
Figure 1 : Key elements of the operant chamber setup.

A Swiss-Webster mouse with a monopolar electrode (1) permanently implanted into the medial forebrain bundle (MFB) is used as an example. A flexible wire lead (2) plugs into the electrode and connects the animal to the stimulator (not shown) through a commutator (not shown) that swivels, enabling the animal to move freely. Mice respond at a wheel (3), since they often do not engage in high rates of lever-pressing for sustained periods. Each one quarter turn of the wheel earns brain stimulation reward (BSR) and illuminates a house light (4), which is wired through an interface (5) to indicate that the electrical circuit is intact. Testing occurs in a standard operant chamber with a clear front (6). A hole for the flexible lead is cut into the center of top of the chamber (not visible), and a groove is cut from the hole to the front of the chamber (not visible) to facilitate handling the animals before and after testing. A 1-cc syringe is taped to the top of the box and the plunger is extended (7) to prevent the flexible lead from entering the groove during testing, which can restrict the movement of the animal. A ring-stand (8) and standard clamp are used to suspend the commutator (not shown) above the center of the hole. The operant chamber can be placed in an external cabinet (9) with doors that enable testing under light- and sound-controlled conditions, although this is not essential because the rewarding efficacy of MFB stimulation makes the behavior extraordinarily rigorous.

Full size image (42 KB)

What are the relative strengths of ICSS?

ICSS has several characteristics that distinguish it from other tests often used to study motivation, including food- or drug-reinforced operant (e.g., i.v. self-administration) or Pavlovian (e.g., place conditioning) paradigms. A prominent difference is that BSR is relatively impervious to anxiety or satiation, unlike many natural rewards (food, sexual behavior). Well-trained rats or mice will engage in ICSS for hours or, if allowed, even days, often to the exclusion of every other behavior. As particularly poignant examples, rats choose access to BSR over access to food6 or heat in a subfreezing environment7, despite lethal consequences. The fact that increases in the intensity of rewarding stimulation (dose) generally produce progressive increases in response rates distinguishes ICSS from other operant paradigms; more food, sustained access to sexually receptive mates, or higher doses of drugs of abuse generally produce progressive decreases in response (self-administration) rates. This difference is potentially important because manipulations that induce aversive states (e.g., drug withdrawal) can also reduce operant behavior8. Manipulations that elevate reward and those that elevate aversion produce dissimilar behavioral outputs in ICSS tests5, 9, 10. ICSS is also advantageous because it can be utilized to study the time course of experimental manipulations, such as acute drug effects10, 11 or chronic alterations in gene expression5. However, it is not particularly useful for the studies designed to measure acquisition rates, because rats and mice generally acquire the operant within a minute if the stimulation electrode is properly placed.

How are ICSS tests conducted?

BSR can be delivered through either monopolar or bipolar electrodes. We use monopolar electrodes because they are smaller in diameter than bipolar electrodes, enabling more precision and less damage to the brain. We deliver cathodal current through the electrode, and allow the charge to dissipate through a grounding electrode (screw) implanted in the skull rather than actively extracting it with anodal charge. We use cathodal current because it is more effective than anodal current, enabling the use of lower stimulation parameters12, 13. It is also safer than anodal current, which can damage tissue by emitting metal ions from the electrode13; indeed, anodal current is often used at the end of ICSS studies to create a mark in the brain at the location of the electrode tip. Several ICSS methodologies can be used to quantify BSR, including the curve-shift14, detection threshold15 and autotitration16 variants. We use the curve-shift variant, because it enables us to generate 'response rate–frequency' relationships across a range of stimulation parameters that are analogous to 'dose–effect' functions in pharmacology. We vary the frequency (Hz) rather than the intensity (μA) of the stimulation, because we want to keep constant the population of neurons that is activated. Typically, we test our animals (rats, mice) in a series of fifteen 1-min trials, each at a different stimulation frequency. Each 1-min trial (Fig. 2a) comprises a 5-s period where the animal receives free (priming) trains (1 s−1) of the cathodal (monopolar) stimulation that is available, followed by a 50-s period where the number of responses (lever presses for rats, wheel revolutions for mice) is counted, followed by a 5-s time-out period during which the frequency is lowered in a 0.05-log unit step. Each series of 15 trials is conducted over a descending series of frequencies (often called a 'pass'; see Fig. 3) and produces a 'response rate–frequency' function: animals typically press at maximal rates for high frequencies, intermediate rates for moderate frequencies and negligible rates for low frequencies (Fig. 2b). There are several ways to quantify stimulation 'thresholds', defined here as the minimum frequency of stimulation required to sustain responding at some arbitrary rate. The most intuitive measure is 'half maximum' (M50), which is analogous to an ED50 in pharmacology. However, manipulations that affect response capabilities (e.g., cause increases or decreases in maximal response rates) can cause artificial shifts in this measure. Accordingly, we quantify ICSS thresholds using Theta-0 (T0), which provides an estimate of the theoretical frequency at which the stimulation becomes rewarding (i.e., response rates become >0). T0 is minimally sensitive to treatment-induced alterations in response capabilities17 (Fig. 2c). Use of the curve-shift variant also enables 'scaling' of treatment effects, which facilitates comparisons of how various treatments affect the function of brain reward systems14. Once established, ICSS thresholds can remain remarkably stable (<10% variability) for the life of the animal, enabling longitudinal studies of long-term treatment effects.

Figure 2: The 'curve-shift' variant of intracranial self-stimulation (ICSS).
Figure 2 : The 'curve-shift' variant of intracranial self-stimulation (ICSS).

(a) Schematic of representative stimulation parameters used in the curve-shift variant of the ICSS paradigm. Typically, a trial begins with five stimulation primes (5 s; P1–P5), with each prime consisting of a 500-ms period of noncontingent stimulation administered at a constant current and frequency followed by a 500 ms time out (TO). Immediately following the primes, there is a 50-s period in which an animal is allowed to self-stimulate. When the animal presses the bar or spins the wheel one quarter turn, stimulation (Stim.) is delivered at the appropriate current and frequency for 500 ms. Each cathodal pulse of current is 100 μs; the frequency determines how many pulses are delivered in this period, which varies from trial to trial (X). There is a 500 ms TO following the stimulation pulse during which the animal can press the bar (or spin the wheel), but no additional stimulation is earned. Following the 50-s test period, there is a 5-s TO period during which no stimulation is available, which ends the 60-s trial. At the end of a trial, the current remains constant, but the stimulation frequency is reduced by 0.05 log units, and the next trial begins. (b) Schematic depicting the theoretical functions that relate response rates (presses) to stimulation frequency (Log (Hz)). The rewarding efficacy of stimulation is measured by nonlinear regression in which a theoretical line (blue, dashed) is drawn through the curve at 60, 50, 40, 30 and 20% of the maximum rate of responding. Either the stimulation frequency that maintains half-maximal responding (M50; purple line, dashed) or that reflects the theoretical point at which the stimulation becomes rewarding (T0; where blue dashed line crosses the x-axis) can be used as the brain reward threshold. In this scenario, brain reward thresholds are either decreased (curve 1) (e.g., by increasing current, or following acute administration of a drug of abuse) or increased (curve 3) (e.g., by decreasing current, or during withdrawal from drug of abuse) relative to the control condition (curve 2). (c) Manipulations that affect response rates (e.g., decreasing (curve 1) or increasing (curve 3) the force required to depress the lever) can artificially affect brain reward thresholds. M50 (purple lines, dashed) is subject to greater variation in response to rate changes compared with T0 (point of rise from x-axis).

Full size image (53 KB)

Figure 3: Data collation and analyses.
Figure 3 : Data collation and analyses.

Examples of raw data generated in rate-frequency determinations in which a drug with reward-enhancing effects (a, cocaine)30 or reward-attenuating effects (b, Salvinorin A (SalvA))11 is administered. After animals are trained over several weeks to stably respond to 'passes' of 15 descending frequency 'trials' (frequencies are decreased in 0.05 log units), an experimental manipulation can be performed. To determine the effect of an acute drug treatment or stimulus on intracranial self-stimulation (ICSS) thresholds, animals are placed in the operant chambers and perform three predrug baseline passes. The animals are removed from the operant chambers, injected with drug, and placed back in the chambers where they perform a predetermined number of postdrug passes (e.g., six passes). The number of presses per trial is reported under each 'pass' heading (i.e., Pass 1, Pass 2). At the end of the experiment, animals are removed from the operant chambers and returned to their home cages. Before calculating T0 and M50 thresholds (see Fig. 2), several steps are taken to ensure that the threshold determinations accurately reflect the behavior of the animal. These steps are applied because the analysis depends on two critical data points: the lowest frequency that supports maximum response rates, and the highest frequency at which responding during a trial is zero (0). Obviously, it is crucial that these steps be applied consistently, in an unbiased (hypothesis-free) manner, to all data sets. 1) We discard the first pass of the predrug baseline determinations because we consider it to be a 'warm-up' period, during which responding is typically variable. 2) If an animal fails to press during a trial (see yellow boxes: presses = '0'), but presses at a rate ≥20% of maximum rate in the trial preceding and following the '0', then we would interpolate by taking the average of the trial immediately before and after (e.g., in (a), Pass 3, Baseline Trial 9 would be interpolated to '51'; (a), Pass 6, Postdrug Trial 6 would be interpolated to 77.5; (a), Pass 5, Postdrug Trial 12 would not be changed). 3) Anytime the animal fails to press in the first trial (see yellow box in (b), Pass 2, Postdrug Trial 1), we replace the '0' with a value that is equal to the maximum presses for that pass. After thresholds and maximum rates are determined for each postdrug pass, they are compared back with Baseline values to generate (c) % baseline threshold and (d) % baseline maximum rate. In the case of (a) Pass 1, postdrug, the reward-enhancing effect of cocaine is so pronounced that the animal responds through all frequencies, making threshold a determination for this pass impossible, since there is no '0' (c; nd).

Full size image (77 KB)

How are ICSS data interpreted?

Many manipulations can affect the efficacy of BSR. This is reflected by shifts in the functions that relate response rates to stimulation frequency (Fig. 2b). For simplicity, these shifts are often expressed (or described) as 'percent baseline threshold'. Generally, each day the animal is first tested to establish baseline (pretreatment) thresholds, and then again immediately after the treatment. Alternatively, chronic treatment effects can be compared to a pre-manipulation baseline, since baseline thresholds tend to be remarkably stable over weeks and months in well-trained animals. Treatment-induced leftward shifts in ICSS thresholds (Fig. 2b, green curve) imply that the stimulation is more rewarding as a result of a treatment (reflecting hyperfunction of brain reward systems), whereas rightward shifts (Fig. 2b, red curve) imply that it is less rewarding (reflecting hypofunction of reward systems). Drugs of abuse decrease the amount of stimulation required to sustain responding, indicated by leftward shifts in rate–frequency functions and decreased ICSS thresholds14, 18. Conversely, agents that block drug reward (e.g., dopamine antagonists) increase the amount of stimulation required to sustain responding, indicated by rightward shifts in rate–frequency functions and increased thresholds14. These agents also block the ability of rewarding drugs to cause leftward shifts. Drug withdrawal—which produces the symptoms of major depression in humans—also causes rightward shifts and elevations in ICSS thresholds9, 19, 20. Thus ICSS is sensitive to the manipulations that increase reward, decrease reward (anhedonia) or increase aversion (dysphoria).

It is important to remember BSR is simply one type of reward. When interpreting ICSS data, the conclusions must be inferential and carefully considered within the context of this paradigm. That is, the fact that drugs of abuse such as cocaine reduce ICSS thresholds implies that the rewarding effects of cocaine add to (or facilitate) the rewarding effects of the stimulation, making lower amounts of the stimulation more effective. Thus it should be concluded that cocaine has reward-related (or reward-facilitating) effects—rather than rewarding effects, per se—in this assay. Similarly, the fact that treatments such as drug withdrawal increase ICSS thresholds implies that they have reward-attenuating effects.

Experimental applications of ICSS

Historically, most ICSS studies have focused on the effects of manipulations, including acute drug treatment10, 11, 14, 21, chronic drug treatment21, 22, spontaneous drug withdrawal9, 20, precipitated drug withdrawal23 and lesions24. ICSS is sensitive to the reward-related effects of drugs that have addictive properties in humans, but are not invariably identified in other rodent paradigms (e.g., ethanol, nicotine, phencyclidine)18, 25, making it an ideal primary screen for abuse liability. Recently, we have been using ICSS to examine how molecular manipulations affect BSR and, by extension, the function of brain reward systems. As an example, we have used viral vectors to induce selective alterations in the expression of glutamate receptor subunits in the NAc of rats, which produce profound effects on ICSS5. ICSS has also been successfully utilized in mutant mice, to establish causal relationships between genetics and behavior. We have shown that mice with a defect in the function of their circadian gene CLOCK show mania-like signs: they are not only more sensitive to the reward-related effects of cocaine but also substantially more sensitive to BSR itself, requiring far less stimulation to sustain responding26. Other researchers have shown decreased sensitivity to BSR in dopamine D2 receptor-deficient mice27. An exhaustive review of potential applications for ICSS is beyond the scope of this protocol, hence a small sample of representative studies and their main findings are listed in Table 1. The combined use of ICSS and molecular approaches may yield improved models of psychiatric illness and addiction, ultimately leading to an improved understanding of the biological basis of motivation.




  • Rats or mice (theoretically, any strain of rats or mice can be used)
    Caution Experimenters must follow national and institutional guidelines for the care and use of laboratory animals, including local requirements for surgical anesthetic.
    Critical Use of adult animals (rats = 300 g; mice = 20 g) is recommended, since electrodes that are permanently affixed to the skull can be displaced by bone growth.
  • Dental cement: acrylic for rats (e.g., Ortho-Jet;Lang Dental Manufacturing), nonacrylic for mice (e.g., Cerebond;


  • Stereotaxic instrument (e.g., Kopf, Model 900)
  • Operant chambers (e.g., ENV-007CT fitted with ENV-110M levers for rats, or ENV-307CT fitted with ENV-113AM response wheels for mice; Med Associates)
  • Stimulators (one per chamber; e.g., Med Associates, cat. no. PHM-152)
  • Commutator (one per chamber; e.g., Plastics One, cat. no. SL2C) to enable the animal to rotate freely
  • Computer and software to automate ICSS procedures and perform analyses (e.g., available with Med Associates ICSS equipment bundle)
  • Oscilloscope to monitor stimulator output (e.g., Agilent, cat. no. 54621A)
  • Stereoscope (to finish electrodes) (e.g., StereoZoom4; Leica)
  • Surgical supplies (anesthetic; atropine; antibiotic; antibiotic cream; syringes; instruments (scalpel, hemostats); scalpel blades; sutures; depending upon the institution, sterile supplies and post-operative analgesics may be required)
  • Electrodes (monopolar; Plastics One, cat. no. MS303-1-3) (see EQUIPMENT SETUP)
  • Electrode holder for sterotaxic instrument (e.g., Plastics One, cat. no. MH-300)
  • Pin-vise (e.g., Plastics One, cat. no. DH-1)
  • 1/8-in drill bit (e.g., Plastics One, cat. no. D-56)
  • Stainless-steel screws (1/8-in; e.g., Plastics One, cat. no. 080-1/8)
  • Flexible wire leads: one to connect each animal to a commutator (e.g., Plastics One, cat. no. 305-305), and another to connect the commutator to the stimulator (e.g., Plastics One, cat. no. 305-491)
  • Software (see EQUIPMENT SETUP)

Equipment setup

  • Electrodes (monopolar) Each electrode should be coated with polyamide insulation except at the tip, which is cut straight across (rather than sharpened). The anode should be an uninsulated flexible stainless steel wire. Commercially available electrodes are often designed such that both the insulated electrode (0.25 mm) and the uninsulated anode (0.125 mm) are embedded within a threaded plug, which fits with a screw-down flexible wire lead. With this product, we cut the electrode itself (excluding the threaded plug) to a length of 10 mm for rats and 6 mm for mice.
  • Apparatus There are no specific requirements for the design (size, materials) of the operant conditioning chambers, although the manipulandum (lever for rats, wheel for mice) should be metal and mounted low on one wall (3 cm above the floor for rats, 1 cm for mice) to facilitate high responding rates. The front wall of the chamber should be transparent to enable visual inspection of the animal. The chambers can be placed within sound- and light-proof external cabinets (with peep-hole to enable visual inspection of the animals); this is not necessary to ensure reliable performance, although a consistent approach (no external cabinet, external cabinet doors always open, or external cabinet doors always closed) should be utilized. The stimulators should have the capability to deliver square-wave stimulation over a range of ~10–400 μA, and have a separate output monitor to which an oscilloscope is attached. Sources for these components of the apparatus include commercial suppliers (e.g., Med Associates) or local manufacturers with expertise in materials, electronics and software programming.Even if these main components are purchased as a set from a commercial supplier, other components must be obtained separately. A single-channel commutator (available commercially from specialty suppliers such as Plastics One) is placed between the stimulator and the animal, so that the animal can behave freely without tangling the wire leads. An ordinary ring-stand and clamp (available from suppliers such as Fisher Scientific) can be used to suspend the commutator directly above the center of the operant chamber. The wire leads and insulated stimulating electrodes can be made locally or purchased from specialty suppliers. It is often necessary to remove the insulation from the tip(s) of the electrode; this is best accomplished with a scalpel blade and stereoscope, which facilitates inspection and enables quality control. If the insulation on the sides of the electrode is cracked or otherwise damaged, it should be re-insulated, returned to the manufacturer, or discarded, since leakage of current from areas other than the tip often creates motor effects that can interfere with high rates of responding and/or detract from the rewarding effects of the stimulation.Key elements of the operant chamber setup are depicted in Figure 1, using a Swiss-Webster mouse as an example.
  • Software Commercial suppliers (e.g., Med Associates) often provide software for ICSS training, testing and data analysis. Moreover, they are often willing to help customize these programs. We have designed our own customized programs; the algorithms we use for estimating ICSS thresholds are based on a least-squares line of best fit analysis, as described previously17 (see Fig. 2b).


  1. Preliminary proceduresFollowing arrival of the animals to the colony, leave them to acclimate. Acclimation of 3–7 d is recommended.
  2. SurgeryPerform surgery. There are some minor differences between the surgical procedures used for rats (option A) and mice (option B). These differences involve stereotaxic coordinates of electrode placements, the number of stainless steel screws implanted into the skull and the type of cement used to affix the electrode assembly to the skull.
    1. Rats
      1. Anesthetize each rat (≥300 g) according to institutional requirements, and immobilize it in a stereotaxic instrument.
      2. Implant a monopolar electrode into the MFB at the level of the lateral hypothalamus (LH) (2.8 mm posterior to bregma, 1.7 mm lateral to midline, 7.8 mm below dura)28.
      3. Wrap the anode around two of the four stainless steel screws that are threaded into the skull to anchor the electrode assembly.
      4. Create the electrode assembly by covering the screws and the electrode with acrylic dental cement.
      5. Close the borders of the incision with sutures such that the skin wraps tightly around (rather than over) the electrode assembly.
      6. Cover the incision with topical triple antibiotic cream.
      7. Administer postsurgical analgesics according to institutional regulations.Troubleshooting
    2. Mice
      1. Anesthetize each mouse (≥20 g) according to institutional requirements and immobilize it in a stereotaxic instrument.
      2. Implant a monopolar electrode (same as those used for rats) in the MFB at the level of the LH (1.9 mm posterior to bregma, 0.8 mm lateral to midline, 4.8 mm below dura)29.
      3. Wrap the anode around a single stainless steel screw gently threaded into the skull.
      4. Secure electrodes to the skull with nonacrylic dental cement. The cement must adhere to bone, because of the poor calcification of the mouse skull it is the cement (rather than the screws) that affixes the electrode assembly to the skull.
      5. Close the borders of the incision with sutures such that the skin wraps tightly around (rather than over) the electrode assembly.
      6. Cover the incision with topical triple antibiotic cream.
      7. Administer postsurgical analgesics according to institutional regulations.Troubleshooting
  3. Initial training and determination of minimum effective currentAllow the animals to recover for 1 week without testing. During this time, monitor recovery according to institutional requirements.
  4. After 1 week recovery, place the animals in the operant conditioning chambers and train them on a fixed-ratio-1 schedule (FR1) to respond for brain stimulation as described (for rats, see ref. 10; for mice, see ref. 30).
  5. Each time the rat presses the lever or the mouse turns the wheel 1/4 of a rotation, the computer should deliver a 0.5-s train of square-wave cathodal pulses (0.1-ms pulse duration) at a set frequency of 141 Hz (Fig. 2a). Following delivery of the stimulation, utilize a 0.5-s time out during which additional responses are not reinforced. One daily training session of 60–90 min is recommended 5 d per week to accustom the animals to the physical demands of the ICSS procedure. The stimulation current (~100–300 μA for rats; ~50–150 μA for mice) should be adjusted by the investigator to the lowest value that sustains a reliable rate of responding (≥40 rewards per min) for 3 consecutive days. This is considered the 'minimal current', reflecting sensitivity to the rewarding effects of the stimulation. Once the minimal current is identified for each animal, it is held constant for the remainder of the study.Troubleshooting
  6. Rate–frequency trainingAdapt each animal to brief tests with a descending series of 15 stimulation frequencies at the minimum effective current. Each series (or rate–frequency 'curve') should comprise 1-min test trials at each frequency. For each frequency, allow a 5-s 'priming' phase during which noncontingent stimulation is given, a 50-s test phase during which the number of responses is counted and a 5-s time-out period during which no stimulation is available (Fig. 2a). Lower the stimulation frequency by 10% (0.05 log10 units), and start another trial.
  7. After responding has been evaluated at each of the 15 frequencies, repeat the procedure such that each animal is given six such series per day (90 min of training). During the training procedure, the range of frequencies should be adjusted for each animal so that the highest six to seven frequencies sustain responding; this approach enables comparable sensitivity to treatments that elevate or lower ICSS thresholds.
  8. Quantify ICSS thresholds (the frequency at which the stimulation becomes rewarding) by plotting a least-squares line of best fit across the frequencies that sustain responding at 20, 30, 40, 50 and 60% of the maximum rate. Define threshold as the frequency at which the line intersects the x-axis (T0)17. Start testing with drugs or other manipulations (next step) when mean thresholds vary by <10% for ≥5 consecutive days.Troubleshooting
  9. TestingConduct testing using within-session (option A) or between-session (option B) experimental designs.
    1. Within-session testing
      1. Generate three rate–frequency curves for each animal immediately before treatment on each test day (as described in Step 6). The first curve serves as a warm-up period and should be discarded because it is typically unreliable; average the second and third curves to obtain the baseline (threshold and maximal response rate) parameters.
      2. Treat each animal (e.g., with an i.p. injection of cocaine30 or κ-agonist10), and obtain four to six more 15-min rate–frequency curves (as described in Step 6). Animals can be used repeatedly since there is substantial evidence that repeated drug administration does not cause progressive changes in drug sensitivity (e.g., tolerance or sensitization) in the ICSS assay31. We typically give each dose of drug twice, in ascending and then descending order, to confirm that there are no progressive changes in drug sensitivity.
      3. Repeat Step 9A(i and ii) on the next day, but using vehicle rather than drug to ensure that the animal has recovered from earlier treatments and to minimize the possibility of conditioned drug effects.
      4. Analyze group differences using ANOVA; analyze significant effects further using post-hoc tests (e.g., simple main effects tests, Newman–Keuls' t-tests).?Troubleshooting
    2. Between-session testing
      1. Calculate the mean threshold of the last five training sessions, and use this as the baseline threshold for the remainder of the study.
      2. Perform manipulations (e.g., microinjections of viral vectors through indwelling guide cannulae to alter gene expression in discrete brain areas), and immediately obtain four to eight rate–frequency curves (as described in Step 6).
      3. Repeat Step 9B(ii) on each test day5.
      4. Compare postmanipulation thresholds back to premanipulation baseline thresholds. Analyze group differences over days using ANOVA; analyze significant effects further using post-hoc tests (e.g., simple main effects tests, Newman–Keuls' t-tests).Troubleshooting
  10. HistologyAt the end of the experiments, confirm electrode placements by analyzing cresyl violet-stained sections of paraformaldehyde-fixed brains10.Troubleshooting


Step 1, acclimation: 3–7 d
Step 2, surgery: 1 h, plus 7 d recovery
Steps 3–5, FR-1 training and current adjustment: 60–90 min per d per animal, for 2–3 weeks
Steps 6–8, rate–frequency training: 90 min per day per animal, for 3–4 weeks
Step 9, testing: 90–135 min per day per animal, indefinitely (animals can be tested repeatedly and/or re-used)
Step 10, histology: 30 min per animal for killing and paraformaldehyde perfusion, 2 h per animal for histological analyses



Troubleshooting advice can be found in Table 2. Some technical issues commonly seen in the raw data are described in Figure 3.


Anticipated results

In normal animals, differences in measures of minimal effective current (completion of Step 5) are assumed to reflect animal-to-animal variations in electrode placements, which cannot be controlled under even the most stringent conditions. However, it is possible that manipulations occurring before ICSS studies commence could affect this measure. Indeed, we have shown that mice with a constitutive mutation in the function of the gene CLOCK systematically require substantially lower currents to sustain ICSS. These data indicate elevated sensitivity to BSR, and are consistent with other works suggesting that this mutation elevates sensitivity to numerous types of rewards26. Dopamine D2-receptor knockout mice appear less sensitive to BSR27. Early developmental exposure to methylphenidate in rats does not appear to affect sensitivity to BSR itself, despite the fact that this treatment reduces the rewarding impact of cocaine32, as well as palatable food and sexual behavior33. It is possible that the variability inherent in electrode placements makes it difficult to detect the enduring consequences of these early manipulations unless the effect size is quite large.

Rate–frequency determinations yield large amounts of data that can be collated and expressed in many ways. As one example, baseline thresholds are calculated from pretest data (Fig. 3a and b), to which subsequent drug-treatment data can be compared in time bins that reflect the length of each rate–frequency determination (e.g., a 15-min pass). Drugs of abuse such as cocaine transiently decrease ICSS thresholds, indicating that the drug enhances reward18 (see Fig. 3a and c). In contrast, the κ-opioid agonist salvinorin A (SalvA) transiently increase thresholds11 (see Fig. 3b and c); this is the same effect as is caused by drug-withdrawal and dopamine antagonists18, 19, suggesting that this pattern of effects reflects reduced reward (or increased aversion/dysphoria). If 15-min passes are used, it can be determined that the effects of both cocaine and salvA are maximal immediately after injection and have waned by 60 min (Fig. 3c). Alternatively, these data points could be collapsed into a single data point that represents the mean for the entire test period; this approach is particularly useful when examining the effect of a treatment over days5. Treatment effects on maximal response rates should also be analyzed in parallel, since these data indicate whether manipulations affect the capability to respond. Typically, effects on reward can be dissociated from the effects on response capabilities, for example, cocaine has negligible effects on response capabilities, whereas salvA causes nominal reductions in ICSS rates (Fig. 3d). Changes in response rates do not invalidate or confound the data, although they do require additional consideration. In CLOCK mutant mice, for example, cocaine causes both a decrease in thresholds and an increase in response rates26. This pattern of data suggests that both the stimulant and reward-related effects of cocaine are enhanced by this mutation.



The authors were supported in part by grants from the National Institutes of Health (DA012736 and MH063266 to W.A.C.; DA023094 to E.H.C.) while writing this protocol.

Competing interests statement: 

The authors declare  competing financial interests.



  1. Liebman, J.M. Discriminating between reward and performance: a critical review of intracranial self-stimulation methodology. Neurosci. Biobehav. Rev. 9, 45–72 (1983). | Article |
  2. Bielajew, C. & Shizgal, P. Evidence implicating descending fibers in self-stimulation of the medial forebrain bundle. J. Neurosci. 6, 919–929 (1986). | PubMed | ISI | ChemPort |
  3. Yeomans, J.S. et al. Brain-stimulation reward thresholds raised by an antisense oligonucleotide for the M5 muscarinic receptor infused near dopamine cells. J. Neurosci. 20, 8861–8867 (2000). | PubMed | ISI | ChemPort |
  4. Carlezon, W.A. Jr. & Wise, R.A. Microinjections of phencyclidine (PCP) and related drugs into nucleus accumbens shell potentiate brain stimulation reward. Psychopharmacology (Berl) 128, 413–420 (1996). | Article | PubMed | ChemPort |
  5. Todtenkopf, M.S. et al. Brain reward regulated by glutamate receptor subunits in the nucleus accumbens shell. J. Neurosci. 26, 11665–11669 (2006). | Article | PubMed | ISI | ChemPort |
  6. Routtenberg, A. & Lindy, J. Effects of availability of rewarding septal and hypothalamic stimulation on bar pressing for food under conditions of deprivation. J. Comp. Physiol. Psychol. 60, 150–161 (1965). | Article | PubMed |
  7. Carlisle, H.J. & Snyder, E. The interaction of hypothalamic self-stimulation and temperature regulation. Experientia 26, 1092–1093 (1970). | Article | PubMed | ChemPort |
  8. Stewart, J. & Wise, R.A. Reinstatement of heroin self-administration habits: morphine prompts and naltrexone discourages renewed responding after extinction. Psychopharmacology (Berl) 108, 79–84 (1992). | Article | PubMed | ChemPort |
  9. Markou, A., Hauger, R.L. & Koob, G.F. Desmethylimipramine attenuates cocaine withdrawal in rats. Psychopharmacology (Berl) 109, 305–314 (1992). | Article | PubMed | ChemPort |
  10. Todtenkopf, M.S., Marcus, J.F., Portoghese, P.S. & Carlezon, W.A. Jr. Effects of kappa-opioid ligands on intracranial self-stimulation in rats. Psychopharmacology (Berl) 172, 463–470 (2004). | Article | PubMed | ChemPort |
  11. Carlezon, W.A. Jr. et al. Depressive-like effects of the kappa-opioid receptor agonist Salvinorin A on behavior and neurochemistry in rats.. J. Pharmacol. Exp. Ther. 316, 440–447 (2006). | Article | PubMed | ChemPort |
  12. Gallistel, C.R., Shizgal, P. & Yeomans, J.S. A portrait of the substrate for self-stimulation. Psychol. Rev. 88, 228–273 (1981). | Article | PubMed | ChemPort |
  13. Yeomans, J.S. Principles of brain stimulation (Oxford University Press, New York, 1980).
  14. Gallistel, C.R. & Freyd, G. Quantitative determination of the effects of catecholaminergic agonists and antagonists on the rewarding efficacy of brain stimulation. Pharmacol. Biochem. Behav. 26, 731–741 (1987). | Article | PubMed | ChemPort |
  15. Esposito, R.U. & Kornetsky, C. Morphine lowering of self-stimulation thresholds: lack of tolerance with long-term administration. Science 195, 189–191 (1977). | Article | PubMed | ChemPort |
  16. Gardner, E.L. et al. Facilitation of brain stimulation reward by Delta9-tetrahydrocannabinol. Psychopharmacology (Berl) 96, 142–144 (1988). | Article | PubMed | ChemPort |
  17. Miliaressis, E., Rompré, P.P. & Durivage, A. Psychophysical method for mapping behavioral substrates using a moveable electrode. Brain Res. Bull. 8, 693–701 (1982). | Article | PubMed | ChemPort |
  18. Wise, R.A. Addictive drugs and brain stimulation reward. Annu. Rev. Neurosci. 19, 319–340 (1996). | Article | PubMed | ISI | ChemPort |
  19. Barr, A.M., Markou, A. & Phillips, A.G. A 'crash' course on psychostimulant withdrawal as a model of depression. Trends Pharmacol. Sci. 23, 475–482 (2003). | Article |
  20. Goussakov, I. et al. LTP in the amygdala during cocaine withdrawal. Eur. J. Neurosci. 23, 239–250 (2006). | Article | PubMed |
  21. Tomasiewicz, H.C., Mague, S.D., Cohen, B.M. & Carlezon, W.A. Jr. Behavioral effects of short-term administration of lithium and valproic acid in rats. Brain Res. 1093, 83–94 (2006). | Article | PubMed | ChemPort |
  22. Carlezon, W.A. Jr. & Wise, R.A. Phencyclidine-induced potentiation of brain stimulation reward: acute effects are not altered by repeated administration. Psychopharmacology (Berl) 111, 402–408 (1993). | Article | PubMed | ChemPort |
  23. Liu, J. & Schulteis, G. Brain reward deficits accompany naloxone-precipitated withdrawal from acute opioid dependence. Pharmacol. Biochem. Behav. 79, 101–108 (2004). | Article | PubMed | ChemPort |
  24. Arvanitogiannis, A., Riscaldino, L. & Shizgal, P. Effects of NMDA lesions of the medial basal forebrain on LH and VTA self-stimulation. Physiol. Behav. 65, 805–810 (1999). | Article | PubMed | ChemPort |
  25. Collins, R.J., Weeks, J.R., Cooper, M.M., Good, P.I. & Russell, R.R. Prediction of abuse liability of drugs using IV self-administration by rats. Psychopharmacology 82, 6–13 (1984). | Article | PubMed | ChemPort |
  26. Roybal, K. et al. Mania-like behavior induced by disruption of CLOCK function. Proc. Natl. Acad. Sci. USA 104, 6406–6411 (2007). | Article | PubMed | ChemPort |
  27. Elmer, G.I. et al. Brain stimulation and morphine reward deficits in dopamine D2 receptor-deficient mice. Psychopharmacology (Berl) 182, 33–44 (2005). | Article | PubMed | ChemPort |
  28. Paxinos, G. & Watson, C. The Rat Brain in Stereotaxic Coordinates 2nd edn. (Academic Press, San Diego, CA, 1986).
  29. Paxinos, G. & Franklin, K.B.J. The Mouse Brain in Stereotaxic Coordinates 2nd edn. (Academic Press, San Diego, CA, 2001).
  30. Gilliss, B., Malanga, C.J., Pieper, J.O. & Carlezon, W.A. Jr. Cocaine and SKF-82958 potentiate brain stimulation reward in Swiss-Webster mice. Psychopharmacology (Berl) 163, 238–248 (2002). | Article | PubMed | ChemPort |
  31. Carlezon, W.A. Jr. et al. Repeated exposure to rewarding brain stimulation downregulates GluR1 expression in the ventral tegmental area. Neuropsychopharmacology 25, 234–241 (2001). | Article | PubMed | ChemPort |
  32. Mague, S.D., Andersen, S.L. & Carlezon, W.A. Jr. Early developmental exposure to methylphenidate reduces cocaine-induced potentiation of brain stimulation reward in rats. Biol. Psychiatry 57, 120–125 (2005). | Article | PubMed | ISI | ChemPort |
  33. Bolaños, C.A., Barrot, M., Berton, O., Wallace-Black, D. & Nestler, E.J. Methylphenidate treatment during pre- and periadolescence alters behavioral responses to emotional stimuli at adulthood. Biol. Psychiatry 54, 1317–1329 (2003). | Article | PubMed | ISI | ChemPort |
  1. Behavioral Genetics Laboratory, Department of Psychiatry, Harvard Medical School, McLean Hospital, Belmont, Massachusetts 02478, USA.

Correspondence to: William A Carlezon Jr1 e-mail:


These links to content published by NPG are automatically generated.


Nicotine addiction and the lure of reward

Nature Medicine News and Views (01 Jun 1998)