Other limitations include the paper being somewhat unclear about response latencies as when one examines the figures and can see the LED turning on it looks as though this happens at about 25 milliseconds not the mentioned 12 milliseconds. I think they should include some records of the actual flash latency for repeated trials and potentially use higher frame rates if they are indeed going to conclude that latencies are so fast. It would be difficult to do this using a camera frame rate which is roughly equal to the latency.
The authors should use actual measured LED flash latencies and not software time stamps for performance calculations.
I will prepare another set of animals, and perform another set of experiments.
There, I will:
Use a (probably) National Instruments device to record timings of frame acquisition and trigger generation.
Align the acquired trigger timings and acquired video frames, and calculate more precise latency from frame acquisition to trigger generation.
I have ordered several mice (because of another issue).
I am going to start habituation and surgery around the beginning of November.
Reviewer 1 (part 2 of 3)
> Other limitations include the paper being somewhat unclear about response latencies as when one examines the figures and can see the LED turning on it looks as though this happens at about 25 milliseconds not the mentioned 12 milliseconds. I think they should include some records of the actual flash latency for repeated trials and potentially use higher frame rates if they are indeed going to conclude that latencies are so fast. It would be difficult to do this using a camera frame rate which is roughly equal to the latency.
>
> The authors should use actual measured LED flash latencies and not software time stamps for performance calculations.
I will prepare another set of animals, and perform another set of experiments.
There, I will:
1. Use a (probably) National Instruments device to record timings of frame acquisition and trigger generation.
2. Align the acquired trigger timings and acquired video frames, and calculate more precise latency from frame acquisition to trigger generation.
I have ordered several mice (because of another issue).
I am going to start habituation and surgery around the beginning of November.
Need to update the Materials and Methods, Results, and Figures.
I did experiments, and recorded using Spike2 and CED 1401:
https://gin.g-node.org/larkumlab/RealtimeDLC_Spike2Recordings
The analysis showed that the figures are better than the timestamp-based (I think).
https://gin.g-node.org/larkumlab/RealtimeDLC_PerformanceProfiling/src/master/spike2-profiling
Need to update the Materials and Methods, Results, and Figures.
Keisuke Sehara
Reviewer 1 (part 2 of 3)
I will prepare another set of animals, and perform another set of experiments.
There, I will:
I have ordered several mice (because of another issue). I am going to start habituation and surgery around the beginning of November.
I did experiments, and recorded using Spike2 and CED 1401: https://gin.g-node.org/larkumlab/RealtimeDLC_Spike2Recordings
The analysis showed that the figures are better than the timestamp-based (I think). https://gin.g-node.org/larkumlab/RealtimeDLC_PerformanceProfiling/src/master/spike2-profiling
Need to update the Materials and Methods, Results, and Figures.
updated as with this commit