Tracking time and space: how the brain records memories

/

Source: MacDonald, C. J., & Tonegawa, S. (2021). Crucial role for CA2 inputs in the sequential organization of CA1 time cells supporting memory. Proceedings of the National Academy of Sciences, 118(3), e2020698118. https://doi.org/10.1073/pnas.2020698118

 Join in me in a favorite childhood memory! I’m three: swinging inside a red canvas hammock my parents hung between two trees. This memory has place details, like where it occurred and what my senses experienced. Mingled with this information is a sense of time: how I first got into the hammock, then rocked in a certain rhythm, and finally drifted to sleep. How does the brain store memories like this? Are space and time details saved together, separately, or both? Do different parts of the brain help store these memory components?

Figure 1: Schematic of the mouse behavioral task

Figure 1: Schematic of the mouse behavioral task

These are the questions researchers MacDonald and Tonegawa explored in an exciting scientific paper published in 2021. One area of the brain that is particularly important for making memories of events is the hippocampus (HPC). Prior work provided clues about how individual HPC neurons might help represent place and time details of an experience. For example, there are place cells, found in a part of HPC called CA1, that fire electrical signals when animals cross certain areas, like the part of the yard that contained my hammock. There are also time cells in CA1 that fire during intervals of a structured experience. In this way, a group of time cells might fire one after the other, throughout each second of a hammock swing. The authors of this paper wanted to understand if place and time cells are interdependent and if they receive the similar electrical inputs. Specifically, they were interested in the relative importance of input from CA2, another region of HPC that drives activity in CA1, for time and place cell firing.

To test these ideas, the authors tasked mice to alternate between turning left and right to receive a reward. The mice run through a loop after deciding, returning back to the intersection again. The authors added a treadmill that delays the mouse’s return to the intersection by 10 seconds. While running in place, the mouse CA1 is only storing time information. While running through the left or right maze loops, the mouse CA1 encodes space and time. While mice ran through this maze and on the treadmill, the authors recorded CA1 activity using small bendy electrodes inserted painlessly into HPC through tiny openings in the skull.

The scientists then used a technique called optogenetics to selectively turn off CA2. This involved expressing a protein which, when exposed to yellow light, silences the electrical activity of neurons in CA2. This way, the authors could precisely turn off CA2 activity during treadmill or loop running. Turning off CA2 during time-coding reduced how accurately mice alternated left and right. During time coding with the light is on and CA2 input is lost, the authors noted that time cells fired erratically, sending signals outside their precise interval. Looking at many time cells’ activity (Figure 2A) with CA2 inhibition, where brighter colors means increased firing, there is no longer smooth sequence of time cells firing one after another during the 10 second span of treadmill running. Place cell sequences were not affected (Figure 2B). If you turned off CA2 in my brain when I was three, I might falsely remember my hammock bouncing around between locations in its arc instead of smoothly swinging.  

Figure 2A-B. Turning off CA2 inputs disrupts how time but not place sequences are represented by groups of neurons. Each row is a neuron that fired most at a specific time (A) or place (B), sorted by time (A) or by position in the loop (B) shown in the columns.

Figure 2A-B. Turning off CA2 inputs disrupts how time but not place sequences are represented by groups of neurons. Each row is a neuron that fired most at a specific time (A) or place (B), sorted by time (A) or by position in the loop (B) shown in the columns.

Ultimately, this paper suggests that CA2 is a critical driver of time but not place cell activity. With less CA2 input, time cells fire imprecisely, leading to worse memory performance. This deepens our understanding of how event-based memories are stored by HPC, and how our brains record our experiences.

Edited by Manasi Iyer