Objectives Using the filled-duration illusion, this study investigated the existence of an independent temporal code operating in working memory. Extending research suggesting the principle distinction between filled- and empty-temporal-intervals is additional non-temporal information in filled-intervals, we explored whether the working memory mechanism of binding operates between temporal and non-temporal information in filled-interval memory representations. Such a finding would evidence a temporal code independent of non-temporal codes in empty-intervals. Previously no study has applied this aspect of the working memory framework to cognitive levels of time perception. Methods Thirty-two participants completed psychophysical tasks adopting the duration discrimination reminder task procedure in two experiments. One experiment looked at visual, shape information, the other looked at spatial, location information Reference intervals lasted 300msec and test intervals ranged from 200msec – 400msec. Intervals were either filled or empty. In half of the trails test-intervals were marked by non-temporal stimuli not matching those presenting reference-intervals (changing in shape or location). We predicted an effect of unmatched non-temporal information only on filled-intervals. Results Analyses investigated three dependent variables: accuracy, proportion of ‘longer’ responses and bisection point. In the location experiment, a significant effect of unmatched non-temporal stimuli was found only in filled-interval trials, in ‘longer’ analyses. Similar results were reported in the shape experiment, finding an additional effect of interval-type (filled versus empty). Analysis of means showed that unmatched non-temporal information effected filled-intervals in location and shape experiments differently. Conclusion An effect of changes in non-temporal information on only filled-interval trials indicates the existence of an independent temporal code, operating in working memory, with a capability to bind with non-temporal information. We suggest that differences in filled-intervals in the location and shape experiments reflect different binding processes operating between temporal information and these two types of non-temporal information.