The Art-JT system is a real-time pressure monitor controlled by an Arduino microcontroller
The Art-JT system includes three parts: pressure sensor, Analog-to-digital converter (ADC), and Arduino microcontroller board (Fig. 1a). It can be connected with personal computer (PC) and fiber photometry system for online analysis and tagging the jump events. We selected a straingauge type transducer (Fig. 1a-1) as the pressure sensor since mice are lighter in weight and straingauge type transducer can achieve greater sensitivity, temperature compensation and eliminate the interference of non-measured directional stress. We put a chamber (20cmL×20cmW×25cmH) on top of the sensor for Naloxone-precipitated withdrawal jump test. The size of the chamber is based on previous studies to observe the withdrawal reaction in mice (9-12). Considering the matching with the sensor, we chose HX711 as the Analog-to-digital converter (ADC) of this device (Fig. 1a-2). The operating voltage of the pressure sensor is 5-12 V, which can be controlled by the Microcontroller Unit (MCU) direct power supply. The rated range (r) is 3KG, and the measurement accuracy is Class B (0.03%* r =0.1g), which fully meets the requirement of recording the activity of mice. The formula of the relationship between the output voltage U1 of the sensor and the pressure F and the reference voltage U0 is as follows:
U1=U0*RO*F/g/r
(RO is the output of 2 mV/V, g is the acceleration of gravity 9.8 m/s2, r is the rated range of 3 Kg, and the estimated value range of U1 is 0-10 mV.)
The last and most important part is Arduino microcontroller board (Fig. 1a-3). The following is the working principle of Art-JT system: it converts the pressure signal into a voltage signal in the cage, and then converts the voltage signal into a digital signal in ADC, and then Arduino records and judges the signal, finally output at the PC and send the TTL signal to fiber photometry system for marking every jump (Fig. 1b). As Ardiuno code (Supplemental Code1) principle showing in the flowing chart in Fig. 1c, the recording begins when “R” is entered. The signals will be updated 10 times per second. Meanwhile, Arduino used the principle of differential filtering to detect the waveform of the jump, when the jump is confirmed, in addition to being recorded by the Ardiuno, TTL signal was sent to the fiber photometry system at the same time. The recording stops when “X” is entered.
The Art-JT system could record the whole process of jump and provide detailed parameters of the jump
We first tested whether Art-JT system could detect the jump of the mouse in it. Fig. 2a showed a representative trace of the pressure changes during the mouse exploring in the chamber. The positive peak means the pressure decreasing which is caused by the mouse’s four legs or forelimbs leaving the cage board. The negative peak means the pressure increasing which usually is caused by the mouse squatting on the board and preparing to jump. Combining with video and recorded pressure trace, we found three types of the jump. The first one is direct jump (Fig. 2a-1), the mouse just jumped to air directly without any preparation, so we can only see a single positive peak or tiny negative peak followed by a big positive peak. The second type is fake jump (Fig. 2a-2), we can only see a negative peak with this type. The mouse just squatted with its hind legs and was preparing to jump. But somehow the mouse did not leave the floor. For the above two situations, algorithmic improvements are made to the code. Finally, only the classic complete jump was included for further analysis. It could be divided into four stages (Fig. 2a-3): Stage 1 is that the mouse just walking or resting on the floor of the cage. It’s four legs all on the board. At this time, the recording trace was showing the real body weight of this mouse. Stage 2 is that the mouse was preparing to jump and squatted with its hind legs. The trace showed a huge negative peak. The amplitude of the peak could increase one third or one-half of the body weight. Stage 3 is that the mouse was hanging in the air or on the top beam of the cage. The trace reached 0. Stage 4 is that the mouse fell back to the floor. The accelerated velocity produced a big negative peak on the trace. The value could be similar or even bigger than stage 2.
Based on the accurate record of the jumping timing, the body weight of the mouse getting from the Art-JT system, we were wondering whether we could get detailed information from this device. We wrote the codes in Matlab (Supplemental Code 2) to run the further analysis and get various parameters. It included the body weight when the mouse was preparing to jump, squatting force with its hind legs, the latency of the jump, the speed at which the mouse jumps (Velocity) in stage2, the height of the jump, and the number of jumps (N) of the mouse (Fig. 2b). We could evaluate the whole view of the jump from these parameters. For example, the squatting force might reflect muscle strength and also motor system function. The latency and the speed of jumps might demonstrate the level of Naloxone precipitated withdrawal. Besides that, we could get the frequency distribution of the inter-jumping interval (Fig. 2c) and the precise inter-jumping interval during the time course (Fig. 2d) which illustrated the dynamic characteristics of stereotyped jump behavior. With these parameters, we could get more information on jumping behavior which may be helpful to understand its neural mechanisms in the future.
The Art-JT system has accurate number counting and timing recording function for the jump
Although Art-JT system can easily and quickly get the total number of mouse jumps and rich parameters about each jump, we still need to confirm that the jumps we recorded indeed correspond to the real jumps of the mice. So, we compared the data of jumping timing and numbers between observers and our device. Fig. 3a demonstrated the whole traces of the jumping behavior from our device and manual count from one mouse. The top trace is the pressure change curve getting from Art-JT system. The arrow means the timing of a tested mouse putting in the cage. The value increased from 0 to the body weight of the tested mouse. A series of peaks are the jumping behavior of the tested mouse during the whole recording period. The below trace is a timeline getting from manually counting. Both traces showed a similar pattern and also a strong correlation. Then we compared the total number of the jump getting from our system and two independent observers. There is no significant difference between Art-JT system and two observers in the total number of jumps (Repeated ANOVA, F = 2.406, p > 0.05; Fig. 3b). We also examined the jumping number difference between either machine and observer or observer themselves. We found the value of the number difference between observers seems greater than it between the machine and observer, although the statistics did not show the significant difference (One way ANOVA, F = 2.406, p > 0.05; Fig. 3c). By comparing the timing of the jump from manual counting to machine counting, we also noticed that the timing of the jump from manual counting gradually developed a delay compared to machine counting. And the delay time even increased gradually (Fig. 3d) as recording time went by, which may be caused by the tired observer.
The Art-JT system could send the jumping timing to the fiber photometry system for tagging the Ca2+ signal related to the jump
Recently, genetically encoded calcium indicator (GECI), such as GCaMP6, was used for exploring the relationship between the certain neuron populations activities and the certain behavior through in vivo fiber photometry recording. In this method, accurate timing of the behavioral events is important for analyzing the Calcium dynamics correlated the behavior. The Art-JT system could convert the real timing of the jump events to 5V TTL signal and send it to the fiber photometry system as the event tag. To test this possibility, we implanted the optical fiber in the mPFC and M1
(Fig. 4a) in GCaMP6s mice and induced these mice to generate Naloxone-precipitated withdrawal jump. The results showed that the fluctuation of the Ca
2+ signal recorded by the fiber photometry system and the received TTL signal correspond to each other on the time axis in mPFC
(Fig. 4b). The Ca
2+ signal changes in each jump were successfully recorded in these two brain regions and superimposed into a Ca
2+ signal change curve with a small standard deviation
(Fig. 4c). But the pattern of Ca
2+ activity caused by the jump in the two brain regions seems to be different. The Ca
2+ activity in mPFC quickly returned to baseline after the jumping, while the calcium activity in M1 even was lower than the baseline after the jump event stop. And the latency of the peak of Ca
2+ activity in the two brain regions also showed some difference. But statistical analysis was not performed due to the small sample size
(Fig. 4d). This implies there may be an order in which Naloxone-precipitated withdrawal jumps activate each brain region. Moreover, we found there is a difference in the area under the curve (AUC) of Ca
2+ activity between M1 and mPFC between 0-1 seconds (p<0.001)
(Fig. 4e). It suggested that the neuron activities of mPFC and M1 were different in Naloxone-precipitated withdrawal jumps. And our Art-JT system is a powerful tool for exploring the neural mechanism of stereotyped jump behavior.