Table 2.

Performance of the automated head-twitch response (HTR) detection after optimization of the preprocessing procedures.

Experiment	Drug	Age (weeks)	Duration (min)	N	Manual HTR counts	Automated HTR counts	Percent detecteda	False positive detections	Total errorsb	Total error ratec	R value
1	LSD	23	30	24	1,311	1,304	99.47%	2	9	0.69%	0.9995
2	25I-NBOMe	20	30	25	1,472	1,472	100.00%	12	12	0.82%	0.9994
3	2C-I	30	30	27	1,388	1,381	99.50%	5	12	0.86%	0.9998
4	25I-NBMD	45	20	13	297	292	98.32%	0	5	1.68%	0.9988
5	1-Butanoyl-LSD	18	30	31	975	972	99.69%	15	18	1.85%	0.9987
6	DOB-FLY	36	30	23	1,771	1,763	99.55%	14	22	1.24%	0.9997
7	R-(–)-DOI	34	30	8	633	629	99.37%	2	6	0.95%	0.9997
8	Mescaline	37	30	27	855	852	99.65%	2	5	0.58%	0.9999
9	DOB	20	30	33	1,819	1,791	98.46%	17	45	2.47%	0.9991
10	Psilocybin	30	30	16	403	401	99.50%	6	8	1.99%	0.9991
11	2C-T-7	25	30	10	388	387	99.74%	5	6	1.55%	0.9994
Totals:				237	11,312	11,244	99.40%	80	148	1.31%

^a Percent detected = (automated HTR count ÷ manual HTR count) × 100.

^bTotal errors = number of false positive detections + number of head twitches that were detected manually but not by the automated procedures.

^cTotal error rate = total errors ÷ manual HTR count.