29 March 2013. Reducing activity in the thalamus of mice can give rise to cognitive impairments similar to those seen in schizophrenia, reports a study that came out March 21 in Neuron. Led by Christoph Kellendonk of Columbia University, New York, the study paired a designer receptor with a designer drug to selectively quiet neuronal activity in the mediodorsal (MD) thalamus, a region with dense, two-way connections with the prefrontal cortex. This manipulation weakened synchronous activity between MD thalamus and prefrontal cortex, and triggered impairments in cognitive flexibility and working memory reminiscent of those found in schizophrenia.
Click image to see a mini-lecture on this paper at the Cell Press YouTube channel.
The findings argue that problems with the thalamus may be sufficient to induce some of the cognitive deficits encountered in schizophrenia. When it comes to sophisticated mental processes like working memory, the thalamus does not figure as prominently as prefrontal cortex does, but its connections with brain regions far and wide makes it plausibly influential. In schizophrenia, brain imaging has turned up abnormalities in the thalamus (see SRF related news story and SRF news story), including reduced activity in the mediodorsal nucleus of the thalamus (Minzenberg et al., 2009), echoing earlier evidence of structural abnormalities in this region (reviewed in Alelú-Paz and Giménez-Amaya, 2008. But it’s hard to know whether these reflect a cause or consequence of schizophrenia symptoms.
To address this question, the researchers modeled the decrease in mediodorsal nucleus activity in mice using the “designer-receptor exclusively activated by a designer drug” (DREADD) system developed by coauthor Bryan Roth (Armbruster et al., 2007). In this system, a receptor is engineered and introduced to specific regions of the brain, then activated at will with a drug. Specifically, the researchers introduced a mutated human muscarinic receptor (hM4D) into the MD thalamus. This receptor was no longer activated by its usual ligand, acetylcholine, or to any other neurotransmitters; instead, its ligand was clozapine-N-oxide (CNO), an otherwise inert chemical. When activated, the hM4D receptors hyperpolarized neurons by turning on potassium channels. Thus, when the researchers wanted to tone down MD thalamus activity, they injected mice with CNO. Unlike a lesion, this provided a graded manipulation of activity.
Flexibility and memory
First authors Sebastien Parnaudeau and Pia-Kelsey O’Neill began by verifying that their DREADD system worked as planned. The hM4D receptor was expressed exclusively in neurons, including those projecting to the prefrontal cortex. Within minutes, a CNO injection suppressed firing in about half the neurons they recorded. The neurons did not go silent, however, but rather decreased their firing rate by an average of 38.7 percent. These changes were not observed when control mice without the hM4D receptor were given CNO.
The researchers then tested the effects of turning down MD thalamus activity on reversal learning, a measure of cognitive flexibility, which is impaired in schizophrenia. In this paradigm, choosing one stimulus earns a reward, and choosing another gives a penalty. After this pairing is learned, the rule changes so that choosing the stimulus previously associated with a reward gives a penalty, and vice versa. With CNO on board, the hM4D mice could learn the first set of pairings as well as did controls, which included three groups: mice without the hM4D receptor given CNO, mice without the hM4D receptor given saline, and mice with the hM4D receptor given saline. When the pairings were reversed, however, the hM4D receptor-CNO mice didn’t adapt as well as the control groups did, and continued to choose the stimulus previously associated with reward.
To probe working memory, also affected in schizophrenia, the researchers tested the mice in a T-maze task. There, the mice explored the maze to discover which of two arms held a reward; after a delay, their memory for this arm was tested by having them choose the arm that did not originally contain the reward. With CNO on board, the hM4D receptor mice struggled to learn the task, taking about two days longer to reach the same level of correct choices attained by the three control groups. CNO also impaired how well a separate group of mice already trained on the task could show what they remembered: hM4D receptor mice chose the correct arm less often than the control group when given CNO. This suggests that decreased MD thalamus activity mucked with working memory, or the ability to keep information online. Other tests indicated that this manipulation did not interfere with attention, locomotion, and the ability to use spatial information.
Trouble in the beta band
Quieting activity in MD thalamus may have interfered with cognitive flexibility and working memory by disrupting its communication with prefrontal cortex, a region crucial for these mental processes. The researchers assessed the state of this connection by measuring the synchronous activity between MD thalamus and the medial prefrontal cortex (mPFC). As mice performed the T-maze working memory task on which they had already been trained, the researchers found enhanced synchrony in the beta-frequency range, which consists of oscillations between 13-20 Hz, but not in gamma or theta frequencies. Upon treating hM4D mice with CNO, however, this beta-band synchrony weakened, meaning that the MD thalamus and mPFC were often out of step with each other. The decrement in synchrony came with a drop in percentage of correct choices mice made during T-maze performance. For both saline- and CNO-treated animals, a robust correlation between beta-band synchrony strength and accurate choices emerged, which suggests that the pattern of activity flowing between MD thalamus and prefrontal cortex is important for retaining information.
Though beta oscillations are not as well known as theta or gamma oscillations, they, too, have links to cognition, and have been reported as impaired in schizophrenia (Uhlhaas et al., 2006). Although the disrupted beta oscillations and cognitive deficits originated from a subtle manipulation of MD thalamus in this paradigm, it remains to be seen whether the decrease in MD nucleus activity found in schizophrenia reflects a core problem of this region, or is a downstream consequence of a problem elsewhere in the brain. Still, the study suggests that a graded manipulation of a single region can be a fruitful way to explore isolated parts of circuits and their contributions to higher mental processes.—Michele Solis.
Parnaudeau S, O’Neill PK, Bolkan SS, Ward RD, Abbas AI, Roth BL, Balsam P, Gordon JA, Kellendonk C. Inhibition of Mediodorsal Thalamus Disrupts Thalamofrontal Connectivity and Cognition. 2013 March 21. Abstract