Academic journal article The Psychological Record

Scopolamine Effects under a Titrating-Delayed-Nonmatching-to-Position Procedure

Academic journal article The Psychological Record

Scopolamine Effects under a Titrating-Delayed-Nonmatching-to-Position Procedure

Article excerpt

Discrete-trial delayed-matching-to-sample (DMTS) procedures and variants thereof have been used for decades to assess working memory in human beings and other animals (Paule et al., 1998; Pontecorvo and Clissold, 1992). A prototypical DMTS procedure involves presenting one of a number of sample stimuli (e.g., inserting the right or left lever in a rat chamber), requiring a response to that stimulus (e.g., a press of the lever, leading to its withdrawal), presenting two or more comparison stimuli (e.g., the two levers) immediately or after a delay, and rewarding (reinforcing) a response to the stimulus that matches the sample stimulus (e.g. delivering food after a press of the lever that was initially inserted as the sample). Many elaborations of the basic DMTS arrangement are possible. For example, a delayed-nonmatching-to-position (DNMTP) procedure is frequently used to study working memory, as well as the effects of drugs and other perturbations thereon, in rats. In DNMTP as usually arranged, lever location defines the sample stimulus and a response to the lever location not presented as the sample is designated as correct and is reinforced.

DNMTP has several advantages for the study of working memory in rats (Paule et al., 1998). In brief, it has face validity, it is readily automated, and it engenders relatively consistent performance that can be repeated within and across subjects. If an independent variable produces effects on multiple dependent variables (e.g., accuracy during trials with different delays, number of trials completed), the procedure offers the advantage of potentially separating effects of independent variables on memory from general behavioral and motivational impairment. Finally, and importantly, published studies have demonstrated the value of DNMTP for studying the effects of drugs and other variables (Paule et al., 1998).

Despite its strengths, several disadvantages of DNMTP have been noted (Chudasama and Muir, 1997; Han, Pierre-Louis, Scheff, and Robinson, 2000; Nordholm, Moore, and Wenger, 1995). DNMTP has ceiling effects due to high accuracy typically obtained at short delays. When delays are short, accuracy typically approaches 95%-100%; this phenomenon leaves little room for statistically significant improvement. Moreover, performance decrements are hard to detect at long delays, where baseline performance approaches chance and variability characteristically is large. In addition, response biases may develop in subjects at longer delays, especially in the presence of certain drugs, which artificially affect performance accuracy (Nordholm et al., 1995). Employing a nonmatching procedure can reduce such biases (Pache, Sewell, and Spencer, 1999). Also, behavioral effects typically interpreted as "memory impairment" may be due to disruption of behavioral strategies developed to "bridge" the delay interval (Chudasama and Muir, 1997; Paule et al., 1998). Finally, due to the binary nature of trials, Winters and Dunnett (2004) have argued that an arcsine transformation of overall percentage of correct responses is necessary for analysis-of-variance testing. This transformation, which typically reduces the power of analysis-of-variance tests (Han et al., 2000), appears to be rare in published studies.

DMTS procedures, including DNMTP, typically utilize a set of fixed delay values from which a single value is selected pseudo-randomly for each trial (e.g., Winters and Dunnett, 2004). An alternative procedure is to arrange a performance-dependent delay value that is adjusted according to the accuracy of prior responses. For example, Han et al. (2000) arranged delays of 1 s or 11 s for rats exposed to DNMTP. Within sessions, the delay on a given trial was 1 s if the running proportion of correct responses for the 1- and 11-s delays combined was 0.75 or lower. The delay was 11 s if this value was higher than 0.75.

The primary measure of working memory was the proportion of trials at the longer delay. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.