Connections that control defence strategy

When presented with a threat, most mice freeze or hide, but a few respond more aggressively. The brain circuits underlying this behavioural choice have now been unpicked.
Dayu Lin is at the Neuroscience Institute, New York University School of Medicine, New York, New York 10016, USA.

Search for this author in:

My nine-year-old daughter got a pair of mice for Christmas. To acclimatize her new pets to her presence, she put her hands into their cage. One of the mice ran straight to its shelter and hid; the other stayed in a corner, held its tail up against the glass wall and rattled it rapidly. Tail rattling is a common response to stress across species (most famously in snakes1), and is considered a warning to intruders1,2. But why do some mice hide, and others threaten? What brain mechanisms determine which defensive strategy to deploy in the face of potential danger? In a paper in Nature, Salay et al.3 reveal a brain region that could be responsible for this decision in mice.

Salay et al. first surveyed the brain regions that are activated by threatening cues. They placed mice in a test arena, then presented them with an overhead looming stimulus — an expanding dark-coloured circle, to mimic an approaching predator. When the mice detected the looming cues, they either froze or quickly ran towards a shelter under which they could hide from the stimulus. Occasionally, mice rattled their tails once they had reached the safety of the shelter.

To identify the neurons that determine these behavioural responses, the researchers analysed the protein c-Fos, which is expressed rapidly in neurons after they have been highly active. Many brain regions show consistently high levels of c-Fos in response to looming, and one caught the authors’ attention — the ventral midline thalamus (vMT). The vMT is interesting in that it is not a part of the eye-to-brain visual pathway or a motor pathway, as might be expected for neurons involved in processing visual stimuli such as the looming cues in this study. Instead, it receives diffuse inputs from limbic areas and the midbrain (regions that support emotion, motivational behaviours and bodily responses), and projects heavily to higher cognitive areas, such as the prefrontal cortex3. Thus, this region is well suited to signalling the internal state of the animal and guiding its defence strategies.

In vivo recordings of neuronal activity revealed that vMT cells are most active when the animals are responding actively to looming stimuli by running and tail rattling. Salay et al. therefore asked whether vMT activity is sufficient to guide an animal’s defence strategy. The authors injected the vMT with a viral construct carrying a receptor protein that, when activated, excites neurons. The receptor is activated on binding to a small ligand, which the authors administered systemically weeks after injection of the virus. Strikingly, when Salay and colleagues artificially activated vMT cells, the mice no longer hid from the looming stimuli, but instead rattled their tails frequently and ran around the open space. Conversely, when vMT cells were suppressed using a similar approach, the animals spent more time freezing and never rattled their tails in response to looming stimuli.

To investigate how vMT activation affects responses to looming in more detail, the authors genetically modified cells such that they could be activated by a laser, giving millisecond-level control of cell activity. They found that activating vMT cells either in tandem with or 30 seconds before the presentation of looming stimuli promoted the same active defence behaviours, such as tail rattling. This behavioural shift could outlast the stimulation itself, which makes it likely that activating the vMT causes a shift in internal state, instead of eliciting acute motor actions. Consistent with this hypothesis, the researchers showed that vMT activation induces strong and long-lasting autonomic responses indicative of increased arousal, such as pupil dilation, but does not induce changes in motor activity in the absence of looming stimuli.

Next, Salay et al. investigated which projections from the vMT are relevant for guiding defence behaviours. They found that distinct regions of the vMT project to the prefrontal cortex and basolateral amygdala, both of which are implicated in many complex functions, including fear and anxiety. The authors again used a viral technique, this time to manipulate specific vMT outputs, and found that activating vMT projections to the prefrontal cortex pathway promoted an active threat-coping strategy (tail rattling) and increased arousal (Fig. 1). By contrast, activating the vMT–basolateral amygdala projection favoured a different behaviour, freezing. Taken together, the authors conclude that distinct vMT outputs control opposing threat-coping strategies.

Figure 1 | To recede or to challenge. Salay et al.3 report that a region of the brain known as the ventral midline thalamus (vMT) controls how a mouse responds to a threatening visual stimulus, such as an expanding dark circle in a ceiling display that looms overhead. a, The authors provide evidence that signalling from the vMT to the prefrontal cortex (PFC) promotes an active defence strategy, such as making an aggressive tail-rattling motion. b, By contrast, when levels of vMT activation are lower, mice respond to the looming stimulus by hiding.

Escaping and hiding are basic survival instincts when an animal is faced with a powerful predator. However, from time to time, prey animals need to stand their ground. For instance, adult ground squirrels protect pups in a burrow fiercely — after detecting a nearby snake, the squirrel flags its tail and aggressively harasses the snake by biting and kicking dirt towards it. As a result, the snake typically abandons its hunting effort and moves away5. When a weak animal challenges a stronger predator in this way, the animal might be considered ‘brave’. Does it stand to reason, then, that the vMT drives a ‘courageous’ state?

Answers to this question remain unclear, because a direct assessment of courage is not possible in non-human species. Intriguingly, however, the internal state elicited by vMT activation seems to be perceived as positive, because the mice prefer to stay in a chamber in which they can receive the vMT stimulation, given a choice. Future studies that examine the correlation between vMT activity and courage in humans will help us to further understand the emotional state encoded by the vMT.

It is also still unclear whether tail rattling in mice can effectively deter a predator in the wild. But this behaviour did successfully stop my daughter from bothering her pets.

Nature 557, 172-174 (2018)

doi: 10.1038/d41586-018-04747-4
Nature Briefing

Sign up for the daily Nature Briefing email newsletter

Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.

Sign Up


  1. 1.

    Greene, H. W. J. Herpetol. 7, 143–161 (1973).

  2. 2.

    Haber, S. B. & Simmel, E. C. Bull. Psychonom. Soc. 7, 84–86 (1976).

  3. 3.

    Salay, L. D., Ishiko, N. & Huberman, A. D. Nature 557, 183–189 (2018).

  4. 4.

    Van der Werf, Y. D., Witter, M. P. & Groenewegen, H. J. Brain Res. Rev. 39, 107–140 (2002).

  5. 5.

    Clark, R. W. Behav. Ecol. Sociobiol. 59, 258–261 (2005).

Download references