Git1^-/- mice have been generated as described previously [13]. Mice at 2~4 months of age were used for all behavioral assays and LFP/EEG recordings. Experiments were done in accordance with the guidelines of the Animal Welfare Committee of KAIST, Korea.

Amphetamine (U.S. Pharmacopia) and ethosuximide (Tokyo Chemical Industry) were dissolved in 0.9% saline and distilled water to final concentrations of 1.2 g/L and 60 g/L, respectively. Solutions for injection were filtered with Minisart filter (0.2 µm; Sartorius Stedim Biotech). WT and Git1^-/- mice received intraperitoneal injection of amphetamine (4 mg/kg), ethosuximide (200 mg/kg), or the same volume of saline.

Mice were anesthetized by ketamine (Yuhan Corporation). For depth recording, a parylene-coated tungsten electrode (0.005 in, 2 MΩ, A-M Systems, Inc.) was implanted into the ventrobasal or mediodorsal nucleus of the thalamus with grounding electrodes over the cerebellum. LFP recordings were performed 1 week after the implantation. LFP activities (sampled at 200 Hz) of freely moving mice were recorded for 1 h using the NACGather program (Theta Burst Corp.). Amphetamine was delivered intraperitoneally to mice 20 min after basal LFP recording. LFP activity was recorded for 1 h briefly after the injection. LFP recordings were analyzed by Matlab, using EEGLAB and custom-written coding.

The size of the open field box was 40×40×40 cm, and the center zone line was located 6.5 cm apart from the edge. Mice were placed in the center of the chamber in the beginning of the assay, and spontaneous locomotion activity was observed and recorded for 60 min in an open field chamber, briefly after injection. The results were analyzed by Ethovision 3.1 program (Noldus). The total distance moved was obtained by summing the movements made during 10~60 min. All behavioral assays were performed in a blind manner.

The apparatus used in the open field tests was used for object recognition tests. During the sample phase, mice were allowed to explore two identical objects for 10 min. Objects were put in the center of the chamber, and mice were first put in the chamber facing the wall. Exploration time for each object was measured. In the test phase, performed 24 h later, one of the two objects was replaced with a new one, and exploration time for each object was measured. All objects were pre-tested to confirm that there was no difference in object preference using C57BL/6 wild-type mice and C57BL/6-129/SV/Jae hybrid wild-type mice.

As previous studies have identified close connection between ventrobasal thalamus and prefrontal cortex during attention task [15,16], I hypothesized that abnormal theta oscillation in the PFC of Git1^-/- mice might be associated with the abnormal activity of the ventrobasal thalamus.

Thus, I examined the functional properties of the ventrobasal thalamus of Git1^-/- mice by measuring local field potentials (LFPs). Notably, power spectral density analysis of LFP revealed that 3-Hz rhythms were increased and 7~10-Hz rhythms were decreased in Git1^-/- mice (Fig. 1A and B), whereas the high-frequency rhythms (>10 Hz) were unaffected (data not shown). I speculated that these enhanced thalamic 3-Hz rhythms may be associated with the ADHD-like phenotypes observed in Git1^-/- mice [13]. Therefore, I tested whether amphetamine, which has been shown to rescue EEG theta rhythms and hyperactivity in Git1^-/- mice [13], could affect these thalamic rhythms. Indeed, I found that amphetamine normalized 3-Hz rhythms but not the 7~10-Hz rhythms in Git1^-/- mice (Fig. 1C~E). In contrast, amphetamine treatment caused a marked increase in 3-Hz rhythms in wild-type mice (Fig. 1C, D and F). These changes, which are analogous to the opposing effects of amphetamine on hyperactivity in Git1^-/- mice (suppression) and wild-type mice (induction) [13], suggest that 3-Hz thalamic rhythms may be associated with the development of abnormally enhanced theta EEG rhythms and hyperactivity in Git1^-/- mice.

Abnormal theta rhythms in the human cortex are thought to be coupled with those in the thalamus through the thalamocortical pathway, a process that has been termed thalamocortical dysrhythmia (TCD) [17]. To this end, I simultaneously measured cortical EEG rhythms and thalamic activities (LFP) in WT and Git1^-/- mice (Fig. 2A). Indeed, enhanced theta rhythm was observed in the mediodorsal nucleus of thalamus in addition to the cortex (Fig. 2B), and theta rhythms in these two regions were highly synchronized (Fig. 2C). Enhanced coherence was also observed in the gamma range (30~60 Hz), a brain rhythm implicated in attention and memory [18], similar to observations in human TCD [17].

Increased coherence between the Git1^-/- cortex and thalamus in the theta and gamma ranges was markedly reduced by amphetamine (not saline) treatment (Fig. 2D~F). These results collectively suggest that the amphetamine-induced reduction in cortical theta rhythms is associated with the amphetamine-mediated reduction of thalamo-cortical coherent activities. Amphetamine-induced reduction in thalamocortical coherence in Git1^-/- mice began to weaken at ~40 min after amphetamine treatment (Fig. 3).

Inhibitory inputs into the thalamus facilitate hyperpolarization of thalamic neurons, which is required for the recovery of T-type calcium channels from inactivation and low-threshold spikes generation. The latter is a neuronal burst activity closely related to thalamic rhythms that are conveyed to other brain regions, including cortex [19]. Ethosuximide, which is mainly used as an antiepileptic agent, is a well-known T-type calcium channel blocker and suppressant of thalamic bursts [20]. Here, I tested whether ethosuximide could suppress 3-Hz rhythms in the Git1^-/- thalamus.

Similar to the effects of amphetamine, ethosuximide normalized 3-Hz thalamic rhythms in Git1^-/- mice (Fig. 4A and B). In addition, it normalized cortical EEG rhythms in the theta range (3~10 Hz) (Fig. 4C and D). Behaviorally, ethosuximide treatment of Git1^-/- mice rescued hyperactivity in the open-field assay (Fig. 4E and F), but had no effect on the time spent in the center region, which is a measure of anxiety-like behavior (Fig. 4G). These results, together with the amphetamine-mediated rescue of aberrant 3-Hz rhythms, cortical theta EEG rhythms, and hyperactivity [13], strongly suggest that enhanced 3-Hz thalamic rhythms are associated with ADHD-like cortical theta rhythms and hyperactivity in Git1^-/- mice.

Unlike the reported rescue of novel-object recognition memory in Git1^-/- mice by amphetamine [13], ethosuximide did not normalize recognition memory in the novel-object-recognition test (Fig. 4H and I). This result implies that the suppression of enhanced 3-Hz thalamic rhythms is not sufficient to normalize impaired recognition memory in Git1^-/- mice.