This is hardly settled science, but one hypothesis with some empirical support is that
Visual mental imagery and perception share similar cortical representations (Cichy et al., 2012; Kosslyn, 2005; Pylyshyn, 2003). It has been proposed that, while brain forward connections convey information from the outside world, backward connections might have a dominant role during the forming of mental images in the absence of external bottom-up inputs (Ganis and Schendan, 2008; Ishai et al., 2000; Kalkstein et al., 2011; Kosslyn, 2005).
For more theoretical speculation on the neural basis of qualia you could read Orpwood (2017), although by his own admittance, his ideas (at least in the past) were not terribly clearly presented. He does discuss a theoretical example of olfactory qualia. (I'm mentioning this because you said in the comments that you want to go beyond visual ones.)
Also note that some neuroscientists think the problem of qualia is ill-posed, in that it is intrinsically dualist:
Typically, qualia are defined by four characteristics. Qualia are private, meaning they cannot be known unless they are experienced through the first-person perspective. Qualia are intrinsic, which indicates that they are self-sufficient and independent of other items in experience. Qualia are also ineffable, signifying that the experience of qualia cannot be sufficiently conveyed through words alone. Finally, qualia are immediately accessible through consciousness, which means that knowledge of qualia is direct and certain, as opposed to knowledge of the physical world, which is indirect and logically deduced. [...]
This definition, unfortunately, leads us down a murky path towards dualism, a philosophical position which asserts that there are two kinds of real substances: the physical substance out of which the universe is composed, and the non-physical substance out of which the mind is composed.
Since you asked in the past about the private conceptions of word meanings (as a barrier to common sense), that kinda has the same problem.
If we just stick to qualia as brain processes/representation, regarding pain-related words, for example
Language is more than a mere medium when it comes to share our pain experiences. In fact, it has been shown that processing pain-related words is associated with enhanced activation of part of the neural circuitry underlying physical pain experiences. [citing four references in support]
Richter M, Eck J, Straube T, Miltner WH, Weiss T. Do words hurt? Brain activation during the processing of pain-related words. Pain. 2010 Feb;148(2):198–205. pmid:19846255
Richter M, Schroeter C, Puensch T, Straube T, Hecht H, Ritter A, et al. Pain-related and negative semantic priming enhances perceived pain
intensity. Pain Res Manag. 2014 Mar-Apr;19(2):69–74. pmid:24716197
Gu X, Han S. Neural substrates underlying evaluation of pain in actions depicted in words. Behav Brain Res. 2007 Aug 6;181:218–23. pmid:17512615
Osaka N, Osaka M, Morishita M, Kondo H, Fukuyama H. A word expressing affective pain activates the anterior cingulate cortex in the human brain: an fMRI study. Behav Brain Res. 2004 Aug 12;153:123–127. pmid:15219713
I think this is somewhat analogous to the issue of V1 involvement in visual imagery.