Real data analysis

Figure 4.4: The firing behaviors of Neurons 1 and 6. The indirect input (Neuron 6) is considered as very strong and excitatory so that Neuron 1 has a very similar spiking behavior to that of Neurons 6.

Table 4.6: Results of Simulation 4 in threshold spiking neuron model with two extra indirect inputs (see notation in Table 4.1). The associated parameter settings are: λ = 40 Hz, η = 0 Hz, µ = 0.15, σ = 0.1.

n p Method E E₁ E₂ E_3∼5 E^∗ Correct MSPE

100 750 OGA+HDHQ 661 144 11 1 0 817 1.427

OGA+HDHQ+Trim 684 130 3 0 0 817 1.434

OGA+HDBIC 555 10 0 0 0 565 1.955

OGA+HDBIC+Trim 558 0 0 0 0 558 1.980

OGA+BIC 0 0 0 0 875 854 2.616

200 1500 OGA+HDHQ 736 210 3 1 0 950 1.301

OGA+HDHQ+Trim 737 210 3 0 0 950 1.301

OGA+HDBIC 946 1 0 0 0 947 1.290

OGA+HDBIC+Trim 946 1 0 0 0 947 1.290

OGA+BIC 0 0 0 0 957 957 2.191

400 2250 OGA+HDHQ 451 527 20 0 0 998 1.233

OGA+HDHQ+Trim 452 531 15 0 0 998 1.234

OGA+HDBIC 993 4 0 0 0 997 1.241

OGA+HDBIC+Trim 993 4 0 0 0 997 1.241

OGA+BIC 0 0 0 0 998 998 1.950

Experimental setup

• Subjects

All animal procedures were approved by the Institutional Animal Care and Use Com-mittee of National Taiwan University and adhered to the guidelines established by the Codes for Experimental Use of Animals from the Council of Agriculture, Taiwan.

Experiments were conducted on adult female Long-Evans rats weighing 200300 g. At the commence of this study, all subjects were trained to open the gate of a restraining trap. Briefly, rats were acclimated to an arena (50 50 43 cm) for 1 hour each in the first two days. Afterwards rats were fasted overnight and on the next day placed into the arena containing an opened restrainer with rat chow inside. Rats were shaped to attain the skill of gate-opening on the following two days. Rats capable of opening the gate without any help from the experimenter more than ten times in two successive days were defined as an opener. A total of 13 rats completed the experimental task.

• Surgery and recording method

The openers were implanted with stainless steel microwire electrodes into their anterior cingulate cortex (ACC), insular cortex (InC) or primary motor cortex (MI). Two 8-channel microwire array electrodes were implanted in each rat. Briefly, 8 stainless-steel wires individually insulated with Teflon (50 µm OD) were line up linearly with equal inter-electrode distance and a total width of 2.5 mm. Small longitudinal holes were opened in the fronto-parietal bone for implantation into the ACC or InC. The coordinates of the ACC were 1.5−3.5 mm anterior to and 0.6−0.8 mm lateral to the bregma, and 1.6−2.0 mm deep in the cortex. The coordinates of the InC were 1.5−3.5 mm anterior to and 3.0−5.0 mm lateral to the bregma, and 4.5−5.0 mm deep from the surface of the cortex.

The coordinates of the MI were 1−3 mm anterior to and 3 mm lateral to the bregma, and 1.6−2.0 mm deep in the cortex. Once the electrodes of MI were in the target site, electrical stimulation was employed to ascertain their motor fields. Electrical pulses were delivered from a constant current stimulator (AM system, model 2100) consisting of a train of 7 square-wave pulses, each 0.2 ms in duration, 300 Hz in 100 ms train duration.

Intensities of the test electrical stimulation ranged from 30−300 µA. This stimulation evoked movements of muscle in whisker (22.2 %), neck (33.3 %), or upper limb (33.3 %).

No overt body movement could be discerned in the remaining 11% sites of stimulation.

Six rats had implants in the ACC and InC, four rats had implants in ACC and MI, and three rats had implants in InC and MI. A pair of stainless-steel screws (1 mm OD) was placed in the skull bilaterally, 2 mm posterior and 2 mm lateral to bregma for EEG recording. The ground electrode was a stainless-steel screw located over the top of the cerebellum (mid-occipital bone). In addition, several stainless-steel screws were placed in the frontal and parietal bones for anchoring. A pair of seven-stranded stainless-steel wires (793200, A-M systems) was inserted into the neck muscles for EMG recording.

After implantation, the holes in the skull and the implanted electrodes were sealed and secured with dental cement.

• Data acquisition

Neuronal activity was recorded using a Multi-channel Neuronal Acquisition Processor

system (MNAP, Plexon, Dallas, TX). The electrical signals were passed from the head-set to an amplifier and band-pass filtered (spike signals: 154−13 kHz, gain: 10,000- to 32,000-fold; EEG and EMG filters: 0.7−170 Hz, gain: 5000- to 10,000-fold) displayed on an oscilloscope and an audio monitor (Grass AM8). Real-time spike sorting was con-trolled by SortClient (Plexon), and the sampling rate of individual channels was 40 kHz.

Synchronized video signals were acquired through CinePlex (Plexon).

• Experimental protocol

After recovery from the surgery for 5 days, the implanted rat was adapted with the telemetry sensor (TBSI, W016020H07K1A) and habituated in the testing arena for the rescuing task. On the testing day, we recorded behavior by videotape and neuronal ac-tivity by telemetry of the freely moving rat in the arena. In an experimental session, the implanted rat was acclimated to the test arena for 30 minutes and then a restrainer with a trapped rat was placed into the center of the arena. Control sessions included testing an implanted rat with a restrainer either being empty or containing a toy rat. Each type of session contained ten trials and the three types of trials (30 in total) were randomized in order.

• Pro-social behaviors in the rescuing task

Throughout the training and testing sessions, 13 female LE rats completed the 2 experi-ment tasks after implantation of cortical recording electrode arrays. On the testing day of the rescuing task, each rat was presented with 30 trials of the rescuing test includ-ing 10 trials respectively for each of the followinclud-ing three conditions, namely, a restrainer containing another rat, a toy rat, or no rat inside.

Experimental results

A total of 288 single units were recorded from the 13 rats. Among them, 107 units were in the ACC from 10 rats, 82 units in the InC from 9 rats and 82 units in the MI from 7 rats. For each unit, the 5-second spike trains prior to the onset of gate-opening of each trial were aligned and superimposed to obtain a single enhanced 5-second spike train, which is then transformed into a 50-sample time series (through binning with bin size 100 ms.) for causal analysis. For each rat, OGA+HDHQ was used to analyze the causal relationship among all its recorded uints through the time-series signals. The time-lag parameter m was fixed at 5 for all rats and the NSIs were computed once the relevant input and output neurons were identified. Through the method, we have figured out, for each rat and condition, some special groups of neurons (from ACC, InC, or MI) conveyed their neuronal information to some other groups of neurons (from ACC, InC, or MI) before the rat perform the gate-opening act.

In order to investigate the functional significance of the flow information for the se-lected neurons, we examined the correlation between neuronal activity and gate-opening behavior using the latter as a trigger event. The averaging binned values of selected units with congruent flow direction were transformed into Z scores according to the mean and SD values of the baseline period from 5 s to 10 s prior to the onset of gate-opening. We found enhancement of activities in projecting neurons of InC prior to the gate opening

for a rat (Figure 4.5E) but not a restrainer either being empty or containing a toy rat (Figure 4.6E). Such enhancements also found in neurons of MI projected to ACC and InC (Figure 4.5G and 4.5H). Coherently, these changes were absent when the rat was opening the restrainer without a real rat (Figure 4.6G and 4.6H). Hence, we suggest that the opening acts executed by the rat for another rat may conducted by InC and contained the affective component.

From the view of the recipient units, we also found that the activities of the recipient units in the MI were increased after the opening acts and sustained for more than 5 seconds (Figure 4.7G to 4.7I). These enhancements were correlated with the sniffing and social approaching behaviors of the recorded rat toward another relief rat. These results also suggested that the social behaviors of the subject were highly related to the strengthen limbic-motor connections.

Figure 4.5: Average all the projected ACC, InC and MI unit activity changes when the rat was opening the gate for a conspecific. Values in the Y-axis are normalized Z-score calculated from ACC neurons which projected to InC (A, n = 8), MI (B, n = 5), and ACC (C, n = 17) units; InC neurons which projected to ACC (D, n = 5), MI (E, n = 4), and InC (F, n = 15) units; MI neurons which projected to ACC (G, n = 4), InC (H, n = 5), and MI (I, n = 15) units. The red line represents the 99 % confidence interval.

The red arrows above the histograms point to the periods of the activity exceeded mean +2.33 SD with 2 consecutive bins. Bin size = 100 ms.

Figure 4.6: Average all the projected ACC, InC and MI unit activity changes when the rat was opening the gate for a toy rat or nothing inside the box. Values in the Y-axis are normalized Z-score calculated from ACC neurons which projected to InC (A, n = 7), MI (B, n = 10), and ACC (C, n = 4) units; InC neurons which projected to ACC (D, n = 7), MI (E, n = 5), and InC (F, n = 11) units; MI neurons which projected to ACC (G, n = 6), InC (H, n = 2), and MI (I, n = 12) units. The red line represents the 99 % confidence interval. Bin size = 100 ms.

Figure 4.7: Average all the recipient unit activity changes in the ACC, InC and MI when the rat was opening the gate for a conspecific. Values in the Y-axis are normalized Z-score calculated from ACC neurons which received projection from InC (A, n = 5), MI (B, n = 5), and ACC (C, n = 19) units; InC neurons which received projection from ACC (D, n = 9), MI (E, n = 7), and InC (F, n = 18) units; MI neurons which received projection from InC (G, n = 3), ACC (H, n = 6), and MI (I, n = 18) units. The red line represents the 99 % confidence interval. Bin size = 100 ms.

Figure 4.8: Average all the recipient unit activity changes in the ACC, InC and MI when the rat was opening the gate for a toy rat or nothing inside the box. Values in the Y-axis are normalized Z-score calculated from ACC neurons which received projection from InC (A, n = 8), MI (B, n = 6), and ACC (C, n = 8) units; InC neurons which received projection from ACC (D, n = 7), MI (E, n = 2), and InC (F, n = 14) units; MI neurons which received projection from InC (G, n = 4), ACC (H, n = 9), and MI (I, n = 12) units. The red line represents the 99 % confidence interval. Bin size = 100 ms.