2AFC Test Research Articles

VISUAL HAZE DETECTION THRESHOLD ASSESSMENT BY ASCENDING METHOD OF LIMITS AND A TRANSFORMED STAIRCASE PROCEDURE IN APPLE JUICE SIMULATIONS UNDER “ STORE LIKE” CONDITIONS

ABSTRACTVisual haze detection thresholds were assessed in thirty panelists by two methods: the ascending method of limits and a transformed staircase procedure (UDTR, from Wetherill and Levitt 1965). The 3AFC test was used as the response procedure for both methods. Experimental samples consisted of an apple juice simulation (water with food coloring). Haze was added to the samples using formazin, and was calibrated to eleven levels ranging from 0.75–13.30NTU. Haze detection thresholds under store‐like conditions (3.55 ± 0.57 [mean ± S.E.] and 3.16 ± 0.50 NTU by the two methods, respectively) were higher than those found previously under more controlled viewing conditions. No differences were found between methods, genders, or order of method presentation (p > 0.05); a practice effect showed a slight improvement in detection thresholds between sessions (p= 0.05). Nonparametric tests confirmed the similarities in the threshold estimates from the two methods. The two methods estimated visual haze detection thresholds equally well. However, the ascending method of limits may be preferred method due to the shorter time involved in administration.

Oral irritation by sodium chloride: sensitization, self-desensitization, and cross-sensitization to capsaicin

Psychophysical methods were used to investigate the irritant sensory properties of concentrated NaCl. The first experiment investigated potential sensitization and desensitization properties. Subjects rated the intensity of the irritation elicited by 10 successive applications of 5 M NaCl on one side of the dorsal surface of the tongue. The mean irritant sensation increased significantly across trials, consistent with sensitization. To test for self- and cross-desensitization effects of unilateral sequential stimulation with NaCl followed by a 10-min rest period, either 5 M NaCl or 10 μM capsaicin was applied bilaterally. In a two-alternative forced-choice (2-AFC) test, subjects indicated which side of the tongue had a stronger irritant sensation. They also rated the intensity of irritation on each side separately. When NaCl was applied bilaterally, the side not previously receiving NaCl was chosen as stronger by a significant majority of subjects and was given significantly higher intensity ratings, consistent with self-desensitization. In contrast, when capsaicin was applied bilaterally, the side that had previously received sequential NaCl was perceived as having a significantly more intense irritation, consistent with cross-sensitization. In a second experiment, the effect of amiloride on NaCl-evoked irritation was studied. One side of the tongue was treated with 1 mM amiloride, after which 5 M NaCl was applied bilaterally and subjects performed the same 2-AFC and rating procedures. Since amiloride significantly reduced the intensity of the irritant sensation, the contribution of amiloride-sensitive ionic currents or the Na +/H + exchange pump (NHE) are suggested as possible transduction mechanisms in lingual nociceptors mediating NaCl-evoked oral irritation.

Cross-desensitization of capsaicin-evoked oral irritation by high but not low concentrations of nicotine in human subjects

We previously showed that sequential application of a relatively low nicotine concentration (7.4 mM) to the tongue, followed by a rest period, induced self-desensitization, i.e. a reduction of oral irritation elicited by subsequent nicotine, but not cross-desensitization of capsaicin-evoked irritation. We presently investigated if cross-desensitization of capsaicin-evoked irritation might be induced by higher concentrations of nicotine. Nicotine (74 or 300 mM) was applied once, unilaterally to the tongue of human subjects. When the irritant sensation had completely subsided, capsaicin (33 μM) was applied bilaterally. In a 2-alternative forced-choice (2-AFC) test, subjects indicated which side of the tongue had a stronger irritant sensation. They additionally rated the intensity of irritation on each side separately. Pretreatment with 74 mM nicotine did not induce cross-desensitization, since no significant difference was observed in intensity ratings or in choice of the treated vs. non-treated side in the 2-AFC. However, 300 mM nicotine did induce cross-desensitization, since a significant majority of subjects chose the non-treated side as having a stronger capsaicin-evoked irritant intensity and assigned significantly higher ratings to that side. These psychophysical findings are discussed in terms of possible neural mechanisms of desensitization.

Visualizing the difference between triangle and 3AFC judgements

The Thurstone-Ura model of triangle and 3AFC judgements explains the greater statistical power of 3AFC tests compared with triangular tests, but its mathematical expression is abstract. However, if the subjective differences among samples are represented on isometric co-ordinates, it is possible to express a Thurstone model of the triangle and 3AFC tasks as a bivariate normal distribution divided into segments by straight lines representing optimal decision rules. The triangle and 3AFC have different optimal decision rules and hence different segment shapes, which account for the tests' different relationships between sensory acuity and the probability of making a correct response. The isometric representation makes the same predictions of these relationships as more abstract formulations but is much easier to represent graphically and to explain to non-specialists.

Evaluation of a dereverberation technique with normal and impaired listeners.

Increasing amounts of room reverberation generally reduce speech intelligibility for both normal and sensorineurally impaired listeners--with the latter group suffering proportionately more. We have implemented a two-input dereverberation technique (Allen et al., 1977) in order to evaluate whether intelligibility scores can be improved by this processing. Reverberated and dereverberated (processed) monosyllable, CVC words were presented monaurally to normal and impaired listeners in both a 2AFC test and an open-response test. Preliminary results for both groups indicate that although perceived and measured reverberation time is clearly reduced by processing, mean recognition scores--evaluated over words and phonemes--are not significantly altered.